Jonathan Derque wrote:
Le 20 Aug, � 08:05, Yanis Batura �crivit : ...While proofreading the 1153rd gismu in the Russian translation of gi'uste by Evgeny Sklyanin, I discovered that {jarbu} was referenced TWICE in the reference list of {tcadu}. I checked the English original and discovered that it was referenced there twice, too.... Since the list contains {se tcadu} and {ve tcadu}, I think it was meant to emphasize the fact that jarbu2 and jarbu4 are especially relevant. Anyway, there are such errors in the gi'uste (eg. the reference list of {palta} contains {cakna} twice).
There are a lot of *possible* reasons for such errors in the references to the right. They accreted over time from numerous efforts to annotate and classify the gismu semantics, from dictionary-writing to thesaurus efforts. The original right-hand notes, IIRC, came from Veijo Vilva's hyperlinking of similar gismu - a quasi-thesaurus. The idea of adding non-definitional aids to understanding what was intended by the gismu definition, while keeping the core place-structure (which was used by LogFlash and had a maximum fixed length), was the primary source for additional notes. The adding of keywords for purposes of what is now called the "draft dictionary" led to many new notes.
I always considered the right-hand notes to be notes, and thus not falling under the baseline, and tended to freely change them in response to questions and comments. (I think that the community has since overruled me on this, since I was roundly criticized for making even minor explanatory changes without seeking community approval.)
I always tried to annotate things, however briefly, so people would have a clue as to why they were included in the right-hand notes). The actual text wording was always manual, and I make mistakes.
I can definitely say that there is no place where the same word is repeated without an explicit annotation (as in the case of tcadu) where that repetition is anything other than a mistake.
lojbab To unsubscribe from this list, send mail to lojban-list-request@lojban.org with the subject unsubscribe, or go to http://www.lojban.org/lsg2/, or if you're really stuck, send mail to secretary@lojban.org for help.