Hoi,
A language without a grammar does not express. You cannot express anything
by using OmegaWiki and consequently it is not a language. Your approach of
OmegaWiki is the same and as flawed as the original approach to Wikipedia;
it does not cover enough ground so how can you call it an encyclopaedia. You
say you first have to include everything before you can start using it...
This is indeed one of your "best" approaches.
When you say we only need to first and biggest dozen or so. I take it you
want to include Bangla and exclude Dutch. And you find it acceptable to
exclude Indonesian and Pashto, Swahili and Igbo. What you can expect in
OmegaWiki is that the coverage for some languages is better then in others.
This is to be expected and this is something that may change over time. It
is not surprising that English is best represented in
.
With the OmegaWiki approach, you can add words in translation to practically
any language. It takes one person to add one translation and it will be
there for all to see. It is a known approach, it is the Wiki approach. What
is different is that we are quite happy to accept resources and not enter
everything by hand. We are quite happy to reach out.
When you are talking about "diminished returns", how do you explain this to
the communities of all the WMF projects that you just excluded? Do you
really want to tell them that they do not matter? How do you rationalise
yourself out of that corner?
How are you going to deal with homonyms, how are you going to deal with
false friends?
You approach does not work and is the only the best solution indeed when a
"worse solution is better"
Thanks,
GerardM
On Jan 23, 2008 4:32 PM, Gregory Maxwell <gmaxwell(a)gmail.com> wrote:
On Jan 23, 2008 9:46 AM, Gerard Meijssen
<gerard.meijssen(a)gmail.com>
wrote:
Hoi,
OmegaWiki is not a constructed language. Representing is as such
demonstrates clearly that you do not understand its concepts.
Sure it is. It is a conlang without a grammar but it has a lexicon,
its words are the unique identifiers for the "defined meaning"s. It
intends to create a word for every idea expressed in any language.
The success of this approach as a tool for translation depends on it
being possible to find meanings distinct enough so that distinct words
remain distinct, but general enough that words in different languages
will share the same meaning.
What you
propose with tagging it with namespaces makes it even more clear that
some
more understanding of the subject matter would be
advisable. In the
ISO-639-3 there are only some 7000 languages, the ISO-639-6 will include
over 25.000 linguistic entities.
Yes, there are a lot of languages defined.. which makes many things
hard. This is why OmegaWiki invents a new language with the explicit
goal of its words and ideas being a superset of all other languages,
so that all other languages can be mapped into this singular new
language.
I'm quite aware of how many languages are defined but I am also aware
that the overwhelming majority of them are not either distinct or
popular enough to warrant common's specific attention. The approach
you advocate requires first creating an enormous dictionary of ideas
and mapping them into many languages before it is useful at all.
A simple handling of this issue could be undertaken right away with
little to no software development, no acceptance of the OmegaWiki
concept.
Yes, a simple approach will not hope to cover all possible languages.
But your proposal would hardly cover anything at all: You would have
us all so busy subsidizing the lack of interest in the majority of
your languages that no one but extreme language advocates would be
interested in participating at all.
Instead, mapping and translation could be done on demand and as people
are interested, rather than being forced into accepting your grand
vision and undertaking a lot of work which they don't find
interesting.
(.. and, as you should be aware, the word "namespace" has meaning
outside of mediawiki. I was not saying that multiple tag namespaces
need to be accomplished with actual MW namespaces, language prefixes
on tags automatically defaulted by the software, or some other
mechanism would work.)
[snip]
Your later assumptions are based on the
"best" solution you can come up
with. So please do some real thinking in stead of blowing
smoke<http://en.wiktionary.org/wiki/blow_smoke>
I'm sorry that I do not share your position on the reasonableness of
first creating a conlang to unify "25,000 languages" as a reasonable
first step into making image tagging multi-lingual compatible.
Our goal should to serve people not linguists. Commons does not need
to support 25,000 languages in order to serve all the peoples of the
world.
Any attempt to go beyond the dozen or so languages which could be
achieved simply through user contributed redirection will have
diminishing returns and will ultimately hurt the quality of the
results we provide for the overwhelmingly majority of the world in
favor of increasingly tiny minority groups whom do not have enough
size or involvement to support their own interests on the project.
_______________________________________________
foundation-l mailing list
foundation-l(a)lists.wikimedia.org
Unsubscribe:
http://lists.wikimedia.org/mailman/listinfo/foundation-l