From jkominek@miranda.org Wed Jul 24 17:08:07 2002 Received: with ECARTIS (v1.0.0; list lojban-list); Wed, 24 Jul 2002 17:08:07 -0700 (PDT) Received: from miranda.org ([209.58.150.153] ident=qmailr) by chain.digitalkingdom.org with smtp (Exim 4.05) id 17XWAo-0001VS-00 for lojban-list@lojban.org; Wed, 24 Jul 2002 17:08:06 -0700 Received: (qmail 14067 invoked by uid 534); 25 Jul 2002 00:08:05 -0000 Date: Wed, 24 Jul 2002 18:08:05 -0600 From: Jay F Kominek To: lojban-list@lojban.org Subject: Re: [lojban] Possible custom disctionary software? Message-ID: <20020724180805.A12589@miranda.org> References: <20020724234338.GA24963@piclab.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 1.0.1i In-Reply-To: <20020724234338.GA24963@piclab.com>; from lee@piclab.com on Wed, Jul 24, 2002 at 06:43:38PM -0500 X-archive-position: 263 X-ecartis-version: Ecartis v1.0.0 Sender: lojban-list-bounce@lojban.org Errors-to: lojban-list-bounce@lojban.org X-original-sender: jkominek@miranda.org Precedence: bulk Reply-to: lojban-list@lojban.org X-list: lojban-list On Wed, Jul 24, 2002 at 06:43:38PM -0500, Lee Daniel Crocker wrote: > It occurs to me that a few further customizations would make it > a great tool for collaboratively building a dictionary. Specifically, > instead of saving and editing just a single free-form piece of text > for each article, it could present a form to be filled out with > predicate places, definition, examples, etc. Perhaps special > markup for switching among Lojban and English. Facilities for > "locking" pages that have been completed and approved. This is almost what I have in mind. Except, keeping the wiki-esque portion completely free form, and having users organize it in a fashion similar to the Terminology section of the current Wiki. Special markup would be used, such as: {kesyske:l1(d)} -> "kesyske astronomy (x1 is the science of outer space} ^^ ^---- emit the natural language definition for kesyske ||^ ^ (according to the language of the current wiki page). ||\-+--- parenthesize the contents. (the natural language word) |\------ emit the keyword for the first gismu place, according to | the current pages language \------- emit the _l_ojban word itself. (kesyske) All the emitted text would be hyperlinked to relevent pages for editing stuff. All the editing stuff would present you with a threaded comment system for that definition, and a voting mechanism, so that you could vote for which of the competeing definitions you thought was the best. All while handling definition text in an arbitrary number of different languages (including and especially Lojban, mind you). Locking of content wouldn't be needed, as the votes would be tallied every few seconds, and the pointer to the best definition updated to reflect votes. Users would have to register to edit anything in this system, so votes couldn't be flooded in. (And the weight of your vote would shift slowly, depending on how highly rated your own definitions tend to be rated.) Further, to address a concern lojbab seems to like to raise, (even though I've included this in every design I've mentioned, I think), all changes, comments, definitions, etc will be tagged with a user id, so we can see who has done what. (And no, obviously you don't need a user id just to search and use the database, merely make changes.) One could likely search for 'kominek dictionary' in the mailing list archive to find my previous descriptions of what I intend to implement. Sadly, I recently lost all of my design documents except for the SQL database tables. I posted them at one point, just to demonstrate that doing this properly isn't as trivial as a flat file. -- Jay Kominek Genius may have its limitations, but stupidity is not thus handicapped. -- Elbert Hubbard