[Foundation-l] African Languages Wikipedia Bashing on Slashdot
Jeffrey V. Merkey
jmerkey at wolfmountaingroup.com
Mon Aug 28 05:51:07 UTC 2006
If someone can identify one or two native speakers for each African
Language who are willing to spend a couple of weeks
putting together lexicons for Wikitrans for each target language, along
with language synthesis rules, I can start performing
runs for the target laguages. I will need the following:
rogets thesaurus (open version on my website) lexicon
transposition of english words and phrases for the Cherokee lexicons
into these languages.
I already have the ability to synthesize new words for advanced latin
derived scientific language which will speed the creation
of a full African Language Wikipedia with the various languages. After
the translation runs, then there should be a large enough
body of matrerials to start getting a community around it. Starting at
ground zero with groups of people who need food more than
computers will certainly relegate the project to failure from the very
start.
Most of these African languages are going to have similiar challenges to
Native languages in American in that they will not
have evolved modern words for modern concepts.
Also, while folks I notice talk a lot about solutions, which is a good
thing, we will need someone here to take some action and get
these speakers to contact me and get these lexicons put together. I
will also need to construct a rules database with the AI engine
and about 20 or so articles from the runs taken and retensed and
corrected to teach the AI engine how to reorder text and phrasing
for these languages into a readable form.
If the Foundation wants a good test run of WikiTrans on these languages,
this would be an excellent project to get Wikipedia
converted to these languages rather than waiting years (or maybe never)
to get it done. I just read the Slashdot article and they
are bashing the heck out of us over this program announcement.
We have rosetta's stone to use, let's use it. I need the folks doing
the African langauges program to shoot me an email, and I will
give them instructions and we can get them at least a good starting
point of content that will require 8,000,000+ articles on wikibooks
and wikipedia to be proofread, but its better than starting from 0.
I am not open sourcing the translator at this time, but I will assist in
creation of lexicons, rules, syntax, and grammar databases and parsers
and perform and post translation runs into any of these languages to
provide a starting point for wikipedia. I am about 5 years ahead of the
game
with solid written tools moulded around MediaWiki that already does all
of this. Let's apply them to this program and just get the thing done and
shutup these nay saying folks claiming we cannot pull it off -- we can.
Jeff
More information about the wikimedia-l
mailing list