Patent Application
20050076342
This invention relates to sharing and automatic translation of software applications.
In the field of this invention automatic translation of software applications is known from a variety of sources. For example:
Some Java™ sharing (or at least observation) is provided by JOSIT (Java Observation Simulation Inspection Toolkit—see website at www.openchannelsoftware.org/projects/JOSIT), although this is designed for application monitoring and performance tuning and it does not extend to application sharing.
The JASMINE (Java Application Sharing in Multiuser INteractive Environments) system (see document at web address www.el-saddik.org/jasmine.pdf) provides a level of Java™ application sharing, but does not facilitate translation.
However, these known approaches have a number of disadvantage(s):
1. Translated Windows/Java application product QE (Quality Evaluation) normally requires a product rebuild. Products such as TRADOS (see website at www.trados.com) can provide UI (User Interface) preview, but not a full functional preview.
2. Shared conferences, e.g., using Lotus™ Translation Services for Sametime™ (L TSS) for chat, are able to share only single language (i.e., not translated) versions of presentations.
3. Heretofore it has not been possible to use or demonstrate a new untranslated product in another language.
A need therefore exists for a system and method for application sharing based on automatic translation wherein the abovementioned disadvantage(s) may be alleviated.
In accordance with a first aspect of the present invention there is provided a system for application sharing based on automatic translation as claimed in claim 1.
In accordance with a second aspect of the present invention there is provided a method for application sharing based on automatic translation as claimed in claim 11.
One system and method for application sharing based on automatic translation incorporating the present invention will now be described, by way of example only, with reference to the accompanying drawing(s), in which:
Referring to
Application Sharing
Application sharing technology per se is not new and several commercial systems (such as Sametime™ and VNC—Virtual Network Computing) exist already. The key in the application sharing mechanism 110 is to intercept at as high as possible a level (image reproduction is not sufficient) so that text methods can be intercepted. The key platforms and approaches per platform are:
An important part of the system 100 is implementation of a standard translation interface 220 for fast translations from any source by the translation mechanism 210 (as this is ‘live’, human translation is unlikely to be appropriate, but an exception is discussed below). A single interface should be used to achieve this, regardless of the translation type used. It can be thought of as a ‘translation server’ that provides translation functionality to any of the above platforms.
The translation mechanism can be implemented as translation memory 230, machine translation 240 or a combination 250 of translation memory 230 and machine translation with human translation.
Translation memory 230 has particular application to QE, where a product has been translated, and a review of the translations as seen ‘in-product’ is desired (as discussed below). Access to the particular associated Translation Memory is made available to whichever of the appropriate technologies above is being used.
Machine translation 240 is more likely to be used for ‘gist’ translation of content/product, especially for presentations/demo's.
A combination 250 of translation memory and machine translation with human translation may be used in interactive systems (e.g., forum/tech support above), in which it would make sense to combine Translation Memory with a process for providing additional translation requirements. Initially translations could be provided by Machine Translation, with requirements for ‘real’ translations being sent off to Human Translation for more accurate work. This would be returned in Translation Memory for future use. An important point here is that as this is a dynamic system, once returned (into the Translation Memory) these translations would be used the next time a forum, for example, is accessed.
Thus, by interaction with the application sharing mechanism 110 and the translation mechanism 210 as described above, an application in source language 310 can be shared in a target language 320 between: a Windows™ application 330, a Java™ application 340, a web browser 350 or any other desired client 360.
It will be understood that the system 100 described above, facilitates the following uses.
Product Translation QE
Most product translation is done by translating to a translation memory which then requires a ‘rebuild’ to create a translated product. As web sites are largely composed (rather than raw HTML) this ‘rebuild’ process applies also to web content. As the rebuild is quite often a complicated or server hosted process (e.g., product compilation, or web database update), it can be quite difficult for translators to see translation ‘in-place’. Translation on a shared application basis however requires only the ability to run the application (e.g., log on to a web page) and share it, with the sharing being intercepted by the ‘translation server’.
Automatic Presentation Translation
For e-conferences, there is already support for instant-messaging translation (via LTSS—Lotus™ Translation Services for Sametime™—allowing a ‘Machine Translator’ to be invited to an Instant Messaging conversation, resulting in both the original and translated text being forwarded to the recipient). If presentations/demos could also be translated automatically (via machine translation), it would greatly increase the potential for cross-language presentations. In scenarios where live presentation translation is done (e.g., European Union presentations or technical demo's in many non-English speaking countries), it could also be extremely useful.
Product Demo's/Usage
As most products do not have translated versions available in the early stages, it would be useful to be able to hook up either Machine Translation or Translation Memory (if someone is willing to provide the translations) to display an early version product in another language.
Web Site Translation QE (Local)
As web sites are largely composed (rather than just raw HTML) a ‘rebuild’ process often applies to modifying web content. As the rebuild is quite often a complicated or server hosted process (e.g., site compilation, or web database update), it can be quite difficult for translators to see translation ‘in-place’. Translation on a shared application basis, however, requires only the ability to run the application (e.g., log on to a web page) and share it, with the sharing being allied to the local translation memory.
Web Site Usage (Local)
As above, but for usage rather than just QE. As translation memories are less likely to exist, machine translation is likely to be more appropriate.
Web Application Translation (Server)
This architecture also provides for the server based provision of translated content by web-servers (such as the BabelFish™ facility provided by Alta Vista™). Although technically this might not be considered conventionally as application sharing, it does retain the property that the original source is never modified. In fact in this implementation, no ‘persistent’ version of the translation is ever kept (except in Translation Memory/Machine Translation implementations as described above).
However, the server may usefully be configured to ‘fire’ a separate translated version to an additional browser (thus sharing the application, rather than simply translating to the current browser).
It is will be noted that in all of the above implementations, apart possibly from server based translation, application sharing is used, meaning that the original application is still available for reference. This improves the ability to resolve ambiguities (as opposed to attempting to directly translate the original product/content).
It will further be understood that the examples described above can be extended in the following ways.
1. Use of Context
Clearly in any translation environment, the use of context to improve translation accuracy can help hugely. Especially in the case of QE review using translation memory. A scheme for including context appropriate at the point of translation (e.g., Dialog & Control ID for Windows™ Application sharing), would be a huge benefit.
2. Reverse Translations for the Server
In the case of interactive web applications, translation of incoming data might be useful (i.e., reverse translation). This could help enable a multilingual bulletin board (if control languages were used to minimise ambiguity), or at least tech-support type question submission,
3. Formatting for BiDirectionality
Where possible/relevant, it would make sense to modify formatting information also to provide BiDirectional (e.g., Hebrew/Arabic) layout.
4. On Demand Service for All Platforms
The application sharing capability could be offered as a server based service (e.g., similar to—or even an extension of—Sametime™), that would perform translations from the server but provide plugins (or equivalent) for application sharing.
It will be understood that the system and method for application sharing based on automatic translation described above provides the following advantages.
An application sharing system can be used, with a link to an automated translated system (translation memory, machine translation, or a combination, depending on the usage) such that the remote/copy version of the application appears translated. This remote/copy version can be either local to a single machine or displayed remotely with a system such as Sametime™. It can also apply to web sites, applied locally or from the server.
| Number | Date | Country | Kind |
|---|---|---|---|
| 0322915.0 | Oct 2003 | GB | national |
