The latest on Kamusi (for background, see /m11 ):

UPDATE, OCTOBER 17, 2007:

The Kamusi Project has been transferred to a new organization, with a new Web address. We are almost done configuring the dictionary to run on the new website. As soon as the necessary programming is complete, we will announce the new web address on this page.

We sincerely regret this interruption, and look forward to being back online very soon.

--The Editor

Edited by author 10-13-2007 08:40 PM
For those not aware of it, the Kamusi Online Living Swahili Dictionary has been offline due to an issue between it and Yale University, which hosted it at: http://www.yale.edu/swahili/

Solutions to the situation are under discussion, with the dictionary about to reopen on another site.

The site currently has these messages (Sept. 5 announcement and Oct. 4 update), FYI:

UPDATE, OCTOBER 4, 2007:

The Kamusi Project hopes to be back online shortly, at a new address. We are currently waiting for the Yale Office of the General Counsel to approve our transfer agreement to a new non-profit home at which it will be possible to raise the funds necessary to continue the project. This page will redirect you to the new Kamusi site when all transfer details have been completed.
 

September 5, 2007:
The Internet Living Swahili Dictionary has been taken offline.

The project has been ordered to remove all links to the sites that the project has relied on to raise revenue for project maintenance and improvement. Without these links, the project has no income source and cannot function; the project will have a negative account balance after outstanding debts are paid.

We sincerely regret this interruption. The Kamusi Project usually serves over one million pages per month to users around the world.

We are working to resolve the situation. However, the site will remain offline until the following issues have been resolved:

• can the Kamusi Project raise funds through the sale of clocks and other merchandise?
• can the Kamusi Project acknowledge gifts from major supporters?
• can the Kamusi Project include passive advertising alongside search results and other pages?

The relevant section of University policy is the following:

ITS Appropriate Use Policy 1607.C.2 Use that is inconsistent with Yale's non-profit status. The University is a non-profit, tax-exempt organization and, as such, is subject to specific federal, state, and local laws regarding sources of income, political activities, use of property, and similar matters. As a result, commercial use of IT Systems for non-Yale purposes is generally prohibited, except if specifically authorized and permitted under University conflict-of-interest, outside employment, and other related policies. Prohibited commercial use does not include communications and exchange of data that furthers the University's educational, administrative, research, clinical, and other roles, regardless of whether it has an incidental financial or other benefit to an external organization.

The Kamusi Project was in the past granted authorization to raise funds through certain merchandising links. However, documentation for that authorization is no longer available. Therefore, the site must remain offline until all questions about whether the project has authorization to raise funds through passive revenue generation on the site have been resolved.

With sincere apologies,
The Editor

5 September, 2007

FYI, a note about this board was posted at http://www.scribd.com/doc/215472/Swahili-A...r-Applications-SACA

The current issue of the magazine Multilingual has an article that may be of interest:

Hurskainen, Arvi, "A rule-based environment for Swahili development," MultiLingual, #90 (Vo. 18, Issue 6, Sept. 2007) [see http://www.multilingual.com/ but the article is available in full only to subscribers]

"The development of Swahili Language Manager (SALAMA) — an environment for developing various kinds of applications for manipulating Swahili text — began at the University of Helsinki in 1985. Rather than constructing an application from scratch for each computational task, it is more economical to develop a comprehensive management system for a language. It is also assumed that, in order to develop high-performing applications, the language, in this case the written text, must first be "written out" for computational handling. While the human being is able to mentally analyze the text directly from writing, the computer needs the hidden information that the text contains to be spelled out.

. . ."

(links to more info on SALAMA can be found below, in /m2 )

Hello! Good Site! Thanks you! rremtzuxbyasfm

Mradi wa KiLinux unaotekelezwa na Chuo Kikuu cha Dar es Salaam, umekamilisha ujanibishaji (localisation) wa Programu ya kuvinjarai mtandao ya Firefox. Programu hiyo itazinduliwa rasmi tarehe 30/9/06 na Rais Mstaafu wa awamu ya pili Mh. Mzee Ali Hassan Mwinyi. Uzinduzi utafanyika katika Chuo Kikuu cha Dar es Salaam. KARIBUNI WOTE.
Mradi wa KiLinux ulishakamilisha ujanibishaji wa programu za JamboOpen Office.

Edited by author 09-18-2006 01:20 PM
The University of Nairobi, the Local Language Speech Technology Initiative (LLSTI), and OneWorld/OKN (Open Knowledge Network) combined to develop a text-to-speech (TTS) program, mainly for use in mobile telephone messaging in Kenya.

For more info see the LLSTI site at http://www.llsti.org/ (there is an online demo of the Swahili TTS, but its operation is spotty).

The OKN part of the effort has become a separate local NGO called Mobile4Good http://www.mobile4good.com

Don Osborn
Bisharat.net

Another longstanding project (10+ years) in the forefront of Swahili language use on the internet is the Kamusi Project - Living Swahili Dictionary at http://www.yale.edu/swahili/ (see also mention of its forum for discussing Swahili in /m1).

Don Osborn
Bisharat.net

Edited by author 09-18-2006 12:30 PM
Here is one message with information about Swahili that I received recently from Timo Honkela in response to a general question about MT in African lanuages:

Professor Arvi Hurskainen at University of Helsinki has conducted a lot of research on language technology for Bantu languages and especially Swahili. Here are some links related to that research:

- Helsinki Corpus of Swahili:
  http://www.aakkl.helsinki.fi/cameel/corpus/intro.htm

- SALAMA - Swahili Language Manager
  http://www.njas.helsinki.fi/salama/
  (includes a project on Swahili-to-English Machine Translation)

- PhD thesis by Wanjiku Ng'ang'a:
  Word Sense Disambiguation of Swahili
  http://ethesis.helsinki.fi/julkaisut/hum/aasia/vk/nganga/

The SALAMA project has been underway now for over 20 years.

Don Osborn
Bisharat.net