[prev in list] [next in list] [prev in thread] [next in thread]
List: kde-devel
Subject: Candidate for Google's "Summer of Code" [was: Proposed presentation
From: Kurt Pfeifle <k1pfeifle () gmx ! net>
Date: 2005-06-01 18:32:05
Message-ID: 200506011932.05663.k1pfeifle () gmx ! net
[Download RAW message or body]
I do not want to spam this list.... but: wouldn't the "challenge"
outlined (and emphasized by my mark) be a candidate for the current
Google bounty program? (See http://code.google.com/summerofcode.html)
And couldn't you, Kevin, approach Ubuntu/Mark Shuttleworth about
it too -- after all, they started something similar just recently,
named "Rosetta" (https://launchpad.ubuntu.com/rosetta/)
Cheers,
Kurt
---- original message to the aKademy-team --------------------------------
On Wednesday 01 June 2005 16:50, Kevin Donnelly wrote:
> Hi
>
> I'd be grateful if you could consider the following for presentation at
> aKademy in Málaga in August.
>
>
> ===begins===
>
> Title:
> A free lexical toolset for minority languages
>
> Summary:
> This presentation introduces a set of free lexical tools which will assist
> translation teams (by streamlining the KDE translation process) and users
> (by providing language-checking resources on the desktop).
>
> Abstract:
> One of the great attractions of KDE is the fact that it is available in a
> number of languages - for example, KDE 3.4 was released in 46 languages, and
> there are another 34 languages at various stages of completion on the KDE
> translation site (i18n.kde.org). KDE already possesses impressive tools (eg
> KBabel) to handle the translation process. However, for minority languages
> the translation effort (already substantial, and usually short of
> manpower) can be hampered by a lack of lexical tools - many minority
> languages suffer from a limited or fragmented written tradition, and/or from
> limited computational resources, so such tools may never have been developed.
> Even if they do exist, they may not be useable under a free license, or on
> the Linux platform.
>
> About 5 years ago, Kevin Scannell began to address these issues for Irish
> (also known as Gaelic or Gaeilge). Although Irish has a long written
> tradition, only limited use was being made of the possibilities offered by
> computerised natural language processing to create lexical tools for Irish.
> Prof Scannell's main efforts were devoted to two areas: (1) building a
> computerised corpus of Irish words and phrases, and (2) developing tools
> based on this corpus which would help Irish speakers use their language more
> effectively - these would include items such as a spell-checker, a
> grammar-checker, a thesaurus, and so on.
>
> This work has led over the intervening period to a valuable free software
> toolset which can be ported to other languages, and will be of particular
> interest to minority languages. Prof Scannell's corpus-builder, Crúbadán
> (which collects and sorts text from webpages in the relevant language), is
> now gathering data on almost 180 languages, 137 of them minority languages
> with limited computerised lexical resources
> (http://borel.slu.edu/crubadan/index.html), and his grammar-checker
> (Gramadóir) is currently being ported to 9 languages
> (http://borel.slu.edu/gramadoir/index.html). Although the toolset is mainly
> focussed at present on surface lexical features, the ability to deal with
> deeper syntactic structures is being developed.
>
> This presentation reviews the toolset, and the original Irish implementation,
> and then looks in more detail at the porting work done so far for Welsh. A
> step-by-step overview of the porting process will hopefully encourage more
> minority language teams to devote some of their scarce resources to this -
> although the amount of work involved is non-trivial, it will pay dividends in
> the future in terms of the KDE translation process (more consistent
> translations, done more easily). For example, the tools can be integrated
> into the KDE translation workflow (as they are for Irish) so that updates
> to .po files can be automatically checked for grammar, deprecated
> translations, and so on. Most importantly, it means that when the KDE
> translation does appear, it will be accompanied by the sort of end-user
> materials that will help KDE gain traction in the business and educational
> spheres. The importance of such tools in free desktop adoption is being
> increasingly recognised (for instance, a grammar-checking API has been
> suggested for OpenOffice.org).
>
> The tools can currently be used from the commandline, and via a selection of
> web interfaces. This presentation ends with an invitation (and a challenge!)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> - can the assembled KDE hackers produce, during the course of aKademy 2005, a
> working plugin to allow the use of these tools (particularly the
> grammar-checker, which offers an XML API) from within KWord? This could be
> an additional selling-point for KOffice with users, complementing its clean,
> light structure and the up-and-coming OASIS format.
>
> ===ends===
>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic