Dear all:

I am a member of many mailing lists, and a few other members have mentioned
that
the "Holy Grail" of our community is a full-fledged, full-text search engine
in our field (multilingual terminology).

Many tools exist to meet the needs of the individual translator, the agency,
and the community at large. Top-notch technology is available to create such
a specialized search engine, but the service is yet to be developed. Over
the last few months, I have questioned a number of service providers and
come to the following conclusions.

There are a number of solutions:

1) For the individual (dtSearch <www.dtsearch.com>, Terminotix
<www.terminotix.com>, etc.);
2) For the agency (Terminotix, MultiCorpora <http://www.multicorpora.com/>,
etc.);
3) For the community at large (general search engines).

There are also a number of difficulties involved with building a good
library to index.

1) Finding quality Web resources in many language pairs.

Right now, small collections of URLs are scattered all over the Web, under
different formats, and impossible to consult effectively. Link pages,
mailing list archives, web rings - none of these resources really help us,
unless we have a lot of time on our hands. The Euręka database should be
used as a first-stop portal for thematic research, but contribution has not
been sufficient to reap the full benefits of its unique features.

2) Creating a good taxonomy.

Translators need a search engine that can find resources in a specific
language pair. Domains are also important. Automated taxonomies are
interesting: a number of search appliances (Exalead, Northern Light,
Convera) offer dynamic field recognition. In other words, they produce
search results that are grouped in a number of intuitive categories, based
on keywords. Nevertheless, future search systems are likely to rely on human
input to define each URL's languages in order to generate the most accurate
hits.

3) Financing the search system.

Personal and corporate search appliances are affordable. The price of portal
search solutions, however, is astronomical. First, one must pay a large
subscription fee. Then, each search request is compiled, and a separate bill
is paid by the client (portal or corporation).

Our community is, by nature, a global one. There are about 3,000,000
translators in the world, and I think we can be described as a search-happy
bunch! What is more, most search engines are financed by banner
advertisements and product sales. Since our main interest is the use of a
search engine for professional purposes, we are not heavy buyers. We are
heavy users. How could any specialized portal survive in those
circumstances?

Engineering.com is an example of specialized portal (with search engine)
that is struggling in spite of millions of dollars in revenue. Northern
Light had to discontinue free access to its public search engine. The list
probably goes on.


Is there any hope, one might ask?
---------------------------------

A number of solutions come to mind. A specialized Web portal might not be
able to survive. On the other hand, a general search engine might be
interested in improving its service to the general population and/or its
corporate clients.

In order to build this value-added service, the Euręka database is the best
place to start. Why? Simple. Search-engine providers value a good library of
links. Major directories/search engines are often based, at least in part,
on the ODP (http://dmoz.org/) and, possibly, other human-edited directories.

Without making any partnership proposal, I have discussed this "specialized
search engine" idea with a number of search-engine representatives. One of
them was quite enthusiastic and called it "very interesting".

The main stumbling block is, of course, money. We already have the software
available to build a "translation ODP" (Euręka -
http://www.foreignword.com/Eureka/default.asp). We already have the means to
start sharing links more effectively and preparing for the future.
(Remember, it is all free.)

In order to trigger the interest of a major search engine, the Euręka
database would have to contain many more URLs and grow at a much faster
rate. This interest could manifest itself in a number of ways. For instance:

1) The Google Glossary (http://labs1.google.com/glossary). If it becomes a
standard Google feature, the Google Glossary might be enhanced by a
partnership with the Euręka database. We would then obtain the expensive
search capability we dream of, and Google would attract additional traffic.

2) An existing search engine might also consider adding a category to its
library of links and allow us to search within that category only. Northern
Light is an example of search engine with a category structure. However, it
does not provide a free search service. Portals and corporations pay to
offer NL search capability on their Intranet and on the Web.

3) An existing search engine might be interested in offering a specialized
search service to paying members, for a monthly fee.

Sadly, the Euręka database is too early in its development to be useful for
any search-engine provider. This lack of a truly comprehensive Web
directory, in our field, is what prevents us from moving forward.

Building the Euręka database would allow us to share links and to collect
glossaries, faster and better, on our workstations and shared agency
platforms. For technical and financial reasons, creating a "translation ODP"
is the only way to prepare for a global terminology search engine.
Hopefully, community-search solutions will emerge from Euręka and save us
valuable time.

Marie-Pierre Lessard
English > Canadian French



-----Message d'origine-----
Message : 1 (Euręka Digest / Envoi groupé)
Date : Sun, 05 Jan 2003 13:43:43 -0000
De : "Yvan Cloutier" <ycloutier@...>
Objet : Re : The Holy Grail of Internet searching... What about Eureka?

(...) ŕ part quelques rares axceptions la participation des membres a été
trčs néligeable. (...)