Wiktionary talk:Votes/pl-2011-02/Deprecating less-than symbol in etymologies

As my questions in the previous discussion were unaddressed, I remain unmoved to participate in the poll which I consider to be a classic "this is a stupid divisive issue I feel like forcing a vote on that I admit is pointless". There are few things more contemptible in online communities. - Amgine/^talk 06:59, 12 March 2011 (UTC)Reply

Polling and voting is not evil. Polls and votes can be poorly setup, they can go wrong, but they can also achieve useful results in a legitimate and fair way.

As regards your questions "What are the effects on machine-parsing en.WT content? how might this affect the dbpedia-folx efforts to build ontology parsing of en.WT content? are there similar or developing standards across Wiktionary languages?", which I admit I have not answered: unification of formatting has benefits for anyone trying to parse Wiktionary content. Parsing of etymologies can be made even easier by further use of templating, which the vote does not prevent. In addition, people who are trying to build ontologies don't care about etymologies, only about definitions and semantic relations.

As regards your question "What specific problem does this address?": this is an attempt at unification of formatting of etymologies. Lack of unification is confusing to new contributors of etymologies, as they can go neither by common practice (because it is mixed) nor by policy (because there is none). --Dan Polansky 07:28, 12 March 2011 (UTC)Reply

Polling and voting are by nature divisive, therefore evil in the context of an online community. Discussion leading to a consensus of opinion is preferred, though it is admittedly a slower process.

RDF ontologies of dictionaries would certainly include etymological relations; in fact they could completely obviate etym sections which would, instead, be automatically generated from the meta-data. However, this assumes there is an original and useful etym section with which to begin.

Standardisation, as you point out, is most beneficial for the parsing process. Have you examined the standards used in other Wiktionary languages? In general it's is usual to default to the most-common format across all references, deviating only when there is a demonstrated need to do so.

Presuming all the above, "what specific problem does this address?" You do not make a problem statement which is logical. Users may follow the common, mixed, practice because there is no specific policy, and there is no demonstrated need for a specific policy because, as yet, there is no trans-language standardisation. A possible formation, for which I have no evidence whatsoever, might be Readers are confused by the over-use of symbols and jargon within Wiktionary etymology sections, in particular the use of '<' to indicate descent from an antecedent term. This should be immediately followed by citations of evidence documenting the problem's existence. - Amgine/^talk 20:10, 12 March 2011 (UTC)Reply

Are you going to change your mind about polling and voting? Because the general consensus of opinion is against you. Moreover, it is much easier to create consensus with people who don't label things as evil; labeling things as evil makes it hard for you to conceded the point (as may be needed to form consensus), it gives people the impression you aren't willing to work towards any consensus that isn't your preferred option, and it tends to push other people (who naturally resent their preferred option being labeled with such a strong word) towards a more polarized viewpoint.

In a world where our users were familiar and comfortable with consensus building techniques, it might be the right way to go. But most of us aren't; and consensus can be more divisive than voting. A voter can be on the wrong side, lose, and accept that. In consensus systems, people can't be on the wrong side; if you refuse to come to consensus, you're a troublemaker; in consensus nobody can openly disagree with the system. In a more communal society, that might work; in the individualistic societies most of the editors are from, it's not acceptable to silence nonconformists. Eppur si muove!--Prosfilaes 21:45, 12 March 2011 (UTC)Reply

Well, no, I won't change my mind about polling and voting. Something you should keep in mind is polls require consensus to abide by their outcome, which generally will happen where the poll is perceived to have been fair, but where canvassing, etc. occurs it tends to harm the collaborative nature of the community. The same is true of every consensus model except absolute consensus; one's preferred result may not ultimately be implemented, but where the process is viewed as fair and balanced most will happily support the result - but no side will have 'lost'. - Amgine/^talk 01:30, 13 March 2011 (UTC)Reply

(unindent, responding to Amgine) Discussion leading to a consensus of opinion is often an infinitely slow process, especially if the subject of discussion is a matter of taste: there's no accounting for taste, so there are often very few valid arguments in such a discussion. If you don't believe it, consider how many of the past discussions in Wiktionary have resulted in a consensus and a subsequent action.

Re "RDF ontologies of dictionaries would certainly include etymological relations": I very much doubt that. I have seen some fragments of ontologies, and none of them had anything to do with etymology. If you show me a fragment of an ontology that incorporates etymological information, I may change my mind. Inserted later: Oh, I have misunderstood. Yes, an ontology for a dictionary would model also etymological relations. But I was originally answering to this question: "How might this affect the dbpedia-folx efforts to build ontology parsing of en.WT content?" Within the context of that question: when someone is parsing Wiktionary to build an ontology from it, the etymology sections are irrelevant. But let us drop the subject: even if I concede for the sake of discussion that etymologies are relevant, standardization does not harm efforts to parse Wiktionary etymologies.

I have not examined Wiktionaries in other languages. I have examined other English-language dictionaries, and stated the results in the poll.

I see the lack of unification as a problem. If you think the lack of unification is not a problem at all, here we differ. --Dan Polansky 21:57, 12 March 2011 (UTC)Reply

I don't think arguing about which consensus models are successful is relevant to this discussion.

My earlier questions were primarily open-ended questions which occurred to me when looking over the discussion as topics which had not been addressed. For all intents and purposes they are still not addressed beyond "standardization does not harm efforts to parse Wiktionary etymologies." Since there are already software/XSLT currently parsing wiktionary entries for a range of applications I believe your answer is perhaps flippant. Changes to the way information is presented, especially templates, will have effects on services trying to reuse en.WT's contents. But, as the questions are simply about things I feel are important to think about when making such proposals, there is no need to reply to them more seriously.

Working with a larger vision than any one language is, in my opinion, both the most difficult and a most important goal in decision-making. - Amgine/^talk 01:30, 13 March 2011 (UTC)Reply

(unindent) Again on parsing: the format "from" is already used in many entries in Wiktionary, so any programmatic extraction of etymologies from Wiktionary already needs to deal with that format. Even if, rather hypothetically, a piece of code wanted to deal with the "<" format, this format can be obtained by something like "sed 's/, from / < /g'", or an equivalent code in Java, Python, PHP, Perl, C++ or any other language that comes to mind. My response is in earnest: I am wholy confident that programmatic extraction of etymologies is not negatively affected if this vote passes, and I have explained where my confidence stems from. --Dan Polansky 08:59, 14 March 2011 (UTC)Reply