History: Possible Research Questions

Starting September 2008, Marta Stojanovic and Alain Désilets of the National Research Council of Canada will start a 12 month R&D effort around CLWE.

As many of you know, choosing a good research question is very difficult task, so please help us by reading the possible ideas below, providing comments, and rating them. A good research question is one for which:

• The answer is not known already, and cannot be found easily.
• The answer matters and has large practical consequences for a particular community.

Thx for your help. We are aiming to choosing one of them by mid-september.

BTW: When you rate ideas, make sure you make your own mind and write your answer down before looking at ratings from other folks.

BTW:

Contexts of use

While collaborative translation has applications in a wide range of situation, we are particularly interested in research that will have impact in the following contexts:

• Government organization that have some sort of legal obligation to provide content in multiple languages (ex: Canadian Goverment, UN departments, European Commission departments).
• Companies that need to produce user documentation for their products in multiple languages.

---

Q1: What is the current state of collaborative translationpractices and technologies?

Description

There are lots of sites that are doing collaborative translation, and many technologies that are used to support them. A partial list can be found here:

• Relevant sites and resources

At this point in time, nobody seems to have a good handle on everything that is happening. It would be good to write a good synthesis of what is happening.

For example, we could write a survey that analyzes the different communities and tools in terms of the extent to which they operate without relying on the Assumptions of conventional translation processes.

Why is this question important?

This is important so we know what has been done already, so we can figure out what the important unresolved problems are, and can focus on solving those instead of re-inventing the wheel.

What makes this a research question?

This is not hardcore quantitative research, but it it falls in the category of qualitative research. It will involve gathering information, writing and analysing surveys, and synthesizing the information into a big picture.

Proposal assessment (please prefix your scores with your initials)

Please help us by providing your own assessment of this research question, on three levels.

Importance: To what degree, do you feel that the answer to this question have important concrete consequences for the community of people doing collaborative translation.
1 = Not important, 5 = Critical importance

Workload: How many person month do you think it will take to answer that question?

Research level: To what degree, do you feel that this qualifies as research?
1 = This is not research at all, 5 = This is definitely research.

Make sure you make up your own mind before looking at other people's assessments (you can view them by clicking in the minus sign below).

Your assesment

[+]

Q2: How to best integrate computer-assisted tools into a collaborative translation platform?

Description

Professional translators have all sorts of Computer Assisted Translation (CAT) tools at their disposal (ex: terminology databases, translation memories), which amateur translators working in collaborative fashion often do not have.

Which of these tools should be integrated into collaborative translation platforms, and if so, how?

In this project, we would integrate open source CAT tools into TikiWiki and gather feedback about their usefulness.

Why is this question important?

This is important because CAT tools have great potential for increasing the productivity of volunteer translators in a collaborative environment.

What makes this a research question?

CAT tools are pretty mature, and we know how to build them for professional translators. We also know that they have a good impact on productivity.

But it's not clear to what extend tools need to be different to help amateur translators, and the extent to which it will actually improve their productivity.

Proposal assessment (please prefix your scores with your initials)

Please help us by providing your own assessment of this research question, on three levels.

Importance: To what degree, do you feel that the answer to this question have important concrete consequences for the community of people doing collaborative translation.
1 = Not important, 5 = Critical importance

Workload: How many person month do you think it will take to answer that question?

Research level: To what degree, do you feel that this qualifies as research?
1 = This is not research at all, 5 = This is definitely research.

Make sure you make up your own mind before looking at other people's assessments (you can view them by clicking in the minus sign below).

Your assesment

[+]

Q3: How can Machine Translation help collaborative translation communities?

Description

Collaborative translation communities often do not have sufficient human resources to cover all language pairs, and to provide translation of all content in a timely fashion.

Machine Translation might help in several ways:

• Automatically provide a "gist quality" translation of new content. This would be only a temporary measure until a human finds the time fix it.
• Allow volunteer translators to translate content from a source language that they can't read. For example, the MT system would provide a "bad" English translation of a page written in Japanese, and the user could fix that bad English without having to actually read the original Japanese.

Why is this question important?

This is important because communities don't want to spend most of their human resources and energy in translation as opposed to creation of original content. MT has the potential of providing "good enough" translation at a fraction of the cost in human resources that fully manual translation can offer.

What makes this a research question?

MT is still bleeding edge technology, so application that uses it is definitely research.

This will involve lots of

Proposal assessment (please prefix your scores with your initials)

Please help us by providing your own assessment of this research question, on three levels.

Importance: To what degree, do you feel that the answer to this question have important concrete consequences for the community of people doing collaborative translation.
1 = Not important, 5 = Critical importance

Workload: How many person month do you think it will take to answer that question?

Research level: To what degree, do you feel that this qualifies as research?
1 = This is not research at all, 5 = This is definitely research.

Make sure you make up your own mind before looking at other people's assessments (you can view them by clicking in the minus sign below).

Your assesment

[+]

Q4: How useful is the current implementation of CLWE?

Description

We have made real progress in the CLWE project, thanks to the excellent work by Louis-Philippe Huberdeau. For a demo, see:

• CLWE Demo Screencast

How useful is this to end users as it is now? What are the remaining problems to be addressed?

Why is this question important?

CLWE is still at beta stage, and it is crucial to evaluate it in real-use situations, in order to improve it.

What makes this a research question?

This is not hardcore, quantitative style of research, but it falls within the realm of more qualitative Human Computer Interaction research.

Proposal assessment (please prefix your scores with your initials)

Please help us by providing your own assessment of this research question, on three levels.

Importance: To what degree, do you feel that the answer to this question have important concrete consequences for the community of people doing collaborative translation.
1 = Not important, 5 = Critical importance

Workload: How many person month do you think it will take to answer that question?

Research level: To what degree, do you feel that this qualifies as research?
1 = This is not research at all, 5 = This is definitely research.

Make sure you make up your own mind before looking at other people's assessments (you can view them by clicking in the minus sign below).

Your assesment

[+]