User talk:Connel MacKenzie/Normalization of articles

I think we are soliciting general comments: I don't think SB's comment was directed at anyone. --Connel MacKenzie T C 17:00, 6 May 2006 (UTC)Reply

• Perhaps I misjudged how controversial this is. I can see this as a topic worthy being voted on. But to my understanding, it was. This was a minor skirmish on Wiktionary talk:Entry layout explained, that I though had been resolved.

• SemperBlotto, please explain why you think your second format is better. I don't see it at all. To me, the second format you gave is just inconsistent. --Connel MacKenzie T C 17:00, 6 May 2006 (UTC)Reply

This one is a little bit contentious. As far as I am aware, very few, if any, definitions in Wiktionary are sentences. There is a linguistic principle that a definition should be substitutable for the word it defines, so that in a sentence, the word can be replaced by the definition and leave the meaning of the sentence unchanged. Hence you can replace "I saw a cow" with "I saw an animal with four legs, one at each corner" and the sentence has the same sense (provided cows are the only quadrupeds with a leg at each corner). It is not possible to do this with full sentences: *"I saw a cow is an animal with four legs, one at each corner" is incorrectly formed. Children's dictionaries sometimes use this format to make their definitions easier to understand to youngsters, but it is not appropriate for Wiktionary. Newbies and one-off contributors sometimes post "full-sentence" definitions, and when I see them, I change them to "substitutable" definitions.

Having said that, I think the issue here has been incorrectly worded: I see it as being whether definitions should be all in lower-case or should begin with a capital letter and end in a full stop/period. That is not the same as their being full sentences, because, simply put, they are not. (I'll leave it as an exercise to the reader to look up the rules for what is required for a string of words to be a sentence.)

I have no strong views either way, but agree with the points made about possible confusion between the capitalised and lower-case initial letter of the first word. I probably come down in favour of the lower-case form, for several reasons:

1. Avoidance of the confusion mentioned above.
2. Most other dictionaries do this (although it should be noted that the OED does not). This could be to save ink or space, which, given that the current OED runs to 20 large volumes, is clearly less of an an issue for them.
3. Initial capitals and final full stops suggest sentences, when the definitions are not (although see how I am formatting this list :) ).
4. Consistency with non-English entries, where only translations or glosses are given and these are in lower-case with no final full stop.
5. It makes it clearer that the definition can be substituted for the word.
6. It looks silly when the definition is a single word.
7. Wikification of the initial word is straightforward (] rather than ]).
8. Less effort is required.

So I think that probably puts me firmly against this policy. — Paul G 07:21, 7 May 2006 (UTC)Reply

Wow. Well, spot checking a few Webster's 1913 entries, they do all seem to be in sentence form, not replacement form. Perhaps it is a pondian issue, after all? As far as I knew, the replacement form is only used in smaller dictionaries, to save space.

To be honest, I never understood the replacement form until I read your explanation above, Paul. Thank you for the clarity.

For your reason #4, I was suggesting this rule apply to those as well.

For your reason #6, I am pretty sure all English words can be a sentence by themselves. Subject/noun/verb is of course the correct requirement, but two of the three are and often can be implied. Run. (Yup, that's a sentence.) Fire! (Yup, that's a sentence.) I. (Yup, that's a sentence.) No. (Yup, that's a sentence.) Yes. (Yup, that's a sentence.) Spaghetti. (Yup, that's a sentence.) Leeds. (Yup, that's a sentence.) Fishing. (Yup, that's a sentence.)

For your reason #7, that is a very good point. For nouns, I usually precede the term with "A " to avoid that issue; for verbs I use "To ".

--Connel MacKenzie T C 08:48, 7 May 2006 (UTC)Reply

• I also forgot to mention: the cow example wansn't quite right. Substituting the Wiktionary definition, you'd get "I saw An animal with four legs, one at each corner.." which if you read without regard for capitalization or punctuation, would be coherent, even in your scenario. --Connel MacKenzie T C 08:55, 7 May 2006 (UTC)Reply

I actually don't think this is really part of the normalization process, but something of a different nature. After all, it does make a difference for the reader, doesn't it? So perhaps this has to wait for a second round of discussion, if all this is brought to the BP. But on topic: I'm with Connel here (I think so), and I don't think it's a Pondian issue. I prefer "An animal with four legs, one at each corner." On the other hand, I understand that many definitions are still lacking, and it would be a waste of time to make lacking or embryonic definitions into sentences, as they still need to be expanded. —Vildricianus | t | 09:55, 7 May 2006 (UTC)Reply

I think this has to be taken on an entry-by-entry basis. I feel, with Paul and Stephen, that having to make every definition a full sentence may sometimes be a constraint. However note that with some complicated terms, eg theological terms or abstract concerpts, a ‘replacement form’ is not always possible, and full-sentence (ie natural English) explanation is desirable there. Widsith 10:44, 7 May 2006 (UTC)Reply

I can't see any rationale to support one form over another. The only comment I'd make is that all the definitions for a single word should be formatted in a single form, rather than mixed (which looks bad, as if someone wasn't paying attention to wehat they were doing). --EncycloPetey 14:23, 7 May 2006 (UTC)Reply

I guess my rationale for thinking it was a non-controversial "normalization issue" (which it is not) was that if we are striving for consistency, then all definitions in Wiktionary should appear as sentences. I'd never heard Paul's eloquent rational for the "replacement style" before, and while that does affect my POV, I'm not entirely sold by any means. I still think that to be consistent, Wiktionary should choose a single style...which would force it to be "sentence style." --Connel MacKenzie T C 17:50, 7 May 2006 (UTC)Reply

Wow. Lots of good arguments here, which I haven'd digested all of. Me, I'm strongly in favor of the "replacement style", for precisely the reason Paul described. I'm somewhat in favor of formatting most definitions (even though they be fragments) with initial caps and final periods, if for no other reason than that I think they look better that way. But there are exceptions. Sometimes, for a complicated definition, "replacement style" is unwieldy, and the reader is better served by a full, grammatical sentence. When a definition needs a second (or third) clarifying sentence, those are full sentences even if the first is a fragment. Finally, when a definition is a translation I agree that it's best left unadorned, with neither initial-cap nor final-stop. So I'm guessing -- at least based on my own proclivities -- that the "policy" here would end up involving several decision points and judgement calls, and would not, after all, end up imposing any kind of absolute uniformity, neither decreeing always or never full sentences, nor always or never caps and periods. –scs 14:30, 5 June 2006 (UTC)Reply

19. Agreed. But see #14. — Hippietrail 16:32, 30 April 2006 (UTC)Reply

19. Never allowing multiple blank lines

I agree except before ---- as I described above in point 13.—Stephen 09:16, 2 May 2006 (UTC)Reply

I (somewhat controversially) altered the default monobook CSS to hide the ---- but increase blank space in its place. I think it ought to be default so I've started a topic on the Beer parlour. I think this is more cross-platform that Stephen's solution which in my experience gives different results in different browsers. Please comment. — Hippietrail 20:35, 6 May 2006 (UTC)Reply

Is this resolved by .css now? Can we say "Never allow multiple blank lines" without exception now? --Connel MacKenzie T C 17:01, 6 May 2006 (UTC)Reply

If this has been sorted out, I agree. — Paul G 07:21, 7 May 2006 (UTC)Reply

Disagree in cases where we have inflection lines and images and wikipedia links and... I'm not sure how this should be handled, but multiple blank lines is the only slution I've found for certain cases where all the included items end up overlapping each other. --EncycloPetey 14:23, 7 May 2006 (UTC)Reply

That's considered as bad layout I guess. There should be other solutions such as adjusting margins. —Vildricianus | t | 15:12, 7 May 2006 (UTC)Reply

No Vild, it's not bad layout. EncycloPetey, what I do in those cases is migrate the float=right items around. So, one before the ==English==, one after ===Etymology===, two after ===Noun=== (before the inflection line, with no blank lines) and any others below. Could you please provide an example term or two, so we can all clearly see what you mean? I am pretty sure I still disagree with multiple blank lines even in this circumstance. The worst I've had to do so far, to date, is use {{-}} to help things align...but not multiple blank lines. --Connel MacKenzie T C 17:54, 7 May 2006 (UTC)Reply

Hmm? I think it is (and this is regarding this section's topic, namely two blank lines). What do you think about another point that widely varies between editors: placement of the {{wikipedia}} link? In most instances it is placed under the POS header, but if Ncik's boxy templates are in place, there's overlap between them. BTW, I didn't even know of {{-}}'s existence. Nice thingy. —Vildricianus | t | 19:15, 7 May 2006 (UTC)Reply

20. controversial: Categories to bottom of entire page, not language section

I have no opinion on this.—Stephen 09:16, 2 May 2006 (UTC)Reply

A related issue is categories between headings where a blank line would otherwise go. I've been leaving these untouched so far. I tend to agree that this one is controversial and that takes it beyond the scope of normalization for now. It would need discussion and perhaps voting on the Beer parlour first. — Hippietrail 22:39, 2 May 2006 (UTC)Reply

See WT:BP#Placing of categories. —Vildricianus 17:47, 3 May 2006 (UTC)Reply

So this should be removed from "The List" and added to "deferred topics"? --Connel MacKenzie T C 17:02, 6 May 2006 (UTC)Reply

I disagree - it is often useful to put categories alongside what they apply to. For example, I routintely put ] in the pronunciation section, which makes it clear that this has been considered. Putting it at the bottom can make it look like it has been overlooked.

Another reason for putting categories in the place where there apply is if, for any reason, that content is deleted (perhaps because the content is wrong), it is clear that the category should be deleted as well. If the category is at the bottom, it is very easy to overlook it.

Furthermore, as categories do not add any formatting to the final page (except for being listed at the end), there is no harm in putting categories close to the content they apply to. (Or maybe they do add whitespace - see #26 below.)

Of course, categories embedded in templates cannot be moved. — Paul G 07:21, 7 May 2006 (UTC)Reply

So, you are saying this should be discussed in the "controversial issues" section instead then, right?

Note that many entries already do have all categories moved to the bottom. If you wish to remove a section now, you currently have to check both places; where the category might be relevant, and at the bottom of the entry. And perhaps the end of that language seciton as well. If we decide at some point that categories belong where they are most relevant, who will go back and change the last 71,000 English entries, to comply? --Connel MacKenzie T C 08:33, 7 May 2006 (UTC)Reply

Disagree for non-English. Users editing a single language need to be able to find all categories listed with that language's section. Otherwise, what's the point of being able to edit long pages by section? I don't want to have to hunt for categories that might be at the bottom of the page or might be carried in an included template or might have been dropped somewhere else. I favor placing all categories for non-English languages immediately after the language header. For English, the bottom of the page makes the most sense. --EncycloPetey 14:23, 7 May 2006 (UTC)Reply

Again: you have to "hunt" for them currently anyway. --Connel MacKenzie T C 17:55, 7 May 2006 (UTC)Reply

21. All HTML auto-converted to UTF-8. & --> &

I agree.—Stephen 09:16, 2 May 2006 (UTC)Reply

I agree. — Paul G 07:21, 7 May 2006 (UTC)Reply

I agree. --EncycloPetey 14:31, 7 May 2006 (UTC)Reply

Personally I'd like to agree, because I'm using a fully UTF-8 capable browser now, but that's a pretty recent thing for me. For people who aren't able to (or don't know how to) edit Unicode, HTML entities are preferable. Are we ready to lock such people out?

(Side notes on what I just said: (1) Yes, since most of our entries do use UTF-8 unapologetically, someone who can't handle it is in big trouble already. But I worry -- even though it's a pretty small possibility -- about someone who makes a new entry using HTML character entities, comes back tomorrow to work on it some more, and can't, because the entities have been autoconverted to UTF-8 behind his back. (2) Even for the majority of editors who do have UTF-8 capable browsers, I'm not sure they're all suitably aware of all of the new distinctions which can and sometimes must be made. If you're used to working just in ASCII, or just in 8859-1/Latin1, you've got some surprises in store when you start using Unicode. There are three kinds of double quotes " “ ” and four kinds of apostrophes/single quotes ' ʼ ‘ ’. This ASCII hyphen - is not the same as this en dash – which is not the same as this minus sign − which is not the same as this em dash —. This turned letter ǝ is not the same as this schwa ə (a distinction which I only learned of after observing the difficulties Hippietrail and Paul G had with the rhymes:English:-eɪʃǝn page). And so on. And if you're not aware of these and several other equally subtle distinctions, you can goober things up, as witness the redirect of aren’t to aren't (or, for that matter, from it's to it’s -- everybody caught the inconsistency there, right?). And even if you are aware of the distinctions and your browser does support UTF-8, you may not know how to enter all those strange characters that aren't on your keyboard...) –scs 15:12, 5 June 2006 (UTC)Reply

Well, we already do, by policy. The choice could have gone either way - towards HTML coded for editing or towards UTF-8 for editing. It went towards UTF-8. The recommendation to avoid HTML coding at all costs, is so that Searching works. --Connel MacKenzie T C 15:26, 5 June 2006 (UTC)Reply

Are you talking ahout everywhere, or just in entry names? —scs 16:34, 5 June 2006 (UTC)Reply

22. All HTML tables converted to wikitables

I don't really understand what this means.—Stephen 09:16, 2 May 2006 (UTC)Reply

In the main namespace especially, we should never see <table>, but instead the {| syntax for specifying tables. --Connel MacKenzie T C 17:04, 6 May 2006 (UTC)Reply

I agree. — Paul G 07:21, 7 May 2006 (UTC)Reply

I agree, but we need a good and easy to find resource available explaining the ins-and-outs of tables, including how all our own templates are set up. Most of what I've needed, I've had to pull from existing examples I've come across here and on Wikipedia. --EncycloPetey 14:31, 7 May 2006 (UTC)Reply

Try m:Help:Table and m:Help:Template. —Vildricianus | t | 14:59, 7 May 2006 (UTC)Reply

I've never seen such HTML tables, only in templates. Perhaps it's a better thing to say that in the main namespace, all wikitables should be hidden in templates, notably the top/mid/bottom ones, and preferably, also inflection templates. —Vildricianus | t | 09:57, 7 May 2006 (UTC)Reply

23. All POS (or equivalent sections) must have at least one line starting with "#" before the next heading

I agree.—Stephen 09:16, 2 May 2006 (UTC)Reply

I have a CSS / Javascript idea for those who don't want to see sense numbers for single-sense articles that should overcome any difficulty here. — Hippietrail 22:49, 2 May 2006 (UTC)Reply

I agree. There should not be intrusive headers between any POS header and the content for that header. --EncycloPetey 14:31, 7 May 2006 (UTC)Reply

Confer entry at I love you. --Connel MacKenzie T C 17:04, 6 May 2006 (UTC)Reply

I agree, subject to what Connel says - a few pages exist purely for the purpose of providing translations (see also day after tomorrow). Perhaps #23 can be reworded to take this into consideration. — Paul G 07:21, 7 May 2006 (UTC)Reply

• I think day after tomorrow deserves a definition. Agreed, it's a bit weird for English, but taking into account that in other languages it's often a single word, I'd say we can just as well define the concept. —Vildricianus | t | 10:02, 7 May 2006 (UTC)Reply

I don't see how this wording doesn't already take that into account. I think my pet peeve I was trying to address when I wrote this was this style:

===Noun===
{{en-noun-reg}}

====Alternative spellings====
*], ]

# Noun definition #1.
# Second noun definition.
# Noun definition #3.

In that, I think the intevening "alt spells" is absolutely unacceptable, for several reasons.

1. I have yet to see where an alternate spelling applies to only a single part of speech.

   (Oh, I bet there is one! I think I had an example once, but I forget. —scs 15:18, 5 June 2006 (UTC))Reply

2. There already is a less intrusive place for alternate spellings (before etymology.)
3. Interleaving them makes parsing harder, as any program must assume the definitions "belong" to the alternate spellings heading, instead of the noun heading.
4. Even in the case where it is possible for alternate spellings to exist for a single part of speech, the sub-heading (as all other sub headings) needs to follow the definitions, not precede them.

--Connel MacKenzie T C 08:28, 7 May 2006 (UTC)Reply

Absolutely agree (despite the I love you and day after tomorrow examples.) —scs 15:18, 5 June 2006 (UTC)Reply

24. POS headings (ongoing individual consolidation discussions - sometimes surprisingly controversial.)

I don't understand what this means.—Stephen 09:16, 2 May 2006 (UTC)Reply

Reducing the number of third level headings in en.wikt:, by combining things like "===Pluralish noun===" or "===Noun phrase===" into 'standard headings' like "===Noun===". --Connel MacKenzie T C 13:38, 2 May 2006 (UTC)Reply

I am in favour of this but again I feel it's too controversial for this discussion. I would love to see it fully fought out and decided elsewhere though. — Hippietrail 22:51, 2 May 2006 (UTC)Reply

Broadly in favour, although trimming everything to "Noun", "Verb", etc, may be too extreme. For example, the OED uses "vbl. sb." (verbal substantive ) and "ppl. a." ("participial adjective") for words like "freezing" in "it's freezing cold" and "this food is suitable for freezing" to show that these derive from the verb "to freeze". I think it is useful to retain parts of speech like "verbal noun" and "participial adjective". On the other hand, "noun phrase" (which should probably be "substantive phrase" anyway) can be trimmed to plain "noun" without loss of useful information - a noun is a noun, and you can tell it is a phrase by seeing that it contains more than one word. — Paul G 07:21, 7 May 2006 (UTC)Reply

I think this should be deferred to the "too controversial" list, for now. All print dictionaries that I know of list parts of speech as abbreviations. Spelling out the meaning of the abbreviation is very confusing to the average reader (who just wants a frikkin definition!) My pet peeve is Transitive vs. Intransitive being split into sepatate headings, instead of a single Verb heading. That however, is spectacularly controversial. Because of the embyonic state of Wiktionary today, the focus most often leans towards assisting entry of translations. As Wiktionary matures, the focus will shift towards making entries readable especially to the average reader. If we want to retain this, on this list, we'll have to specifically exclude my controversial Tr/Intr issue, I think.

I find that "Noun phrase" is utterly useless; I completely agree with Paul that "phrase" should be omitted in that case. OTOH, the "Phrase" heading alone/itself is particularly useful for phrasebook entries. The "Idiom" heading is astoundingly useful; even when a particular form of an idiom can be pigeon-holed into a particular part of speech, other similar forms (that redirect to it) often are not the same part of speech/function! I feel strongly that idioms should be listed under the "Idiom" heading, while any part of speech should be listed at the start of a definition line. That, of course, mimics what print dictionaries do. --Connel MacKenzie T C 08:12, 7 May 2006 (UTC)Reply

Agree with HT. Too controversial for this topic, and it has recently been "discussed" again in the BP. Even though personally, I'd love to see, at last, an exhaustive and final list of POS headings compiled, each with considerable archived arguments for future subdual of contestation. But that'll be for the next round. —Vildricianus | t | 10:12, 7 May 2006 (UTC)Reply

This deserves it's own discussion, but it is worth pursuing to see what POS headers are desired, needed, and used. I thought I had a good handle on these until I started creating Latin entries. --EncycloPetey 14:31, 7 May 2006 (UTC)Reply

I strongly agree this must go to the "controversial" list. Part of the actual resolution will have to be devising a list of headers (confer: /todo2) for discussion. --Connel MacKenzie T C 18:02, 7 May 2006 (UTC)Reply

Even though there are aspects which are controversial, I think we could and should do some consolidation now, as part of this normalization. If you look at Patrik Stridvall's header list at http://tools.wikimedia.de/~stridvall/headers.php under English, you see wild variation. We may not have decided (and may never decide) whether to have "Transitive verb" and "Intransitive verb" or just "Verb", but let's get rid of "Transitive Verb" and "Verb, transitive" and all the other inconsequential variations.

• Replace all f, m, c, n and pl with {{f}}, {{m}}, {{c}}, {{n}} and {{p}} respectively. --Connel MacKenzie T C 06:17, 6 May 2006 (UTC)Reply

Agreed - although, to be clear, I think Connel means typing {{m}}, not {{temp|m}}; is that right?

Correct. I meant {{m}}. --Connel MacKenzie T C 08:01, 7 May 2006 (UTC)Reply

Although this makes no difference to the output, it is possible that some time in the future we might (although I wouldn't like to see it) change {{m}} to expand to masculine rather than m. Having used the template for this everywhere would make this very simple and quick to change. We have already seen this happen with {{p}} which has changed from pl to plural (there is a separate discussion of this elsewhere and so it would be off-topic to discussion here whether this is right or not).

I think that conversation is welcome here. As another side note, I think Hippietrail is doing some experimental magic/css with these particular templates as well. --Connel MacKenzie T C 08:01, 7 May 2006 (UTC)Reply

• Yes, HT made it possible to choose whether they have a period or not: m vs. m.
• No; {{p}} is still pl
• Also, their output is not the same: m has text displaying on hover, which doesn't happen with m
• On topic: can't this be done by a bot? —Vildricianus | t | 10:18, 7 May 2006 (UTC)Reply

• NO, IT CANNOT BE DONE BY 'BOT! I had this particular change in my monobook.js for a long time; it worked correctly about 70% of the time - but there are a gazillion cases where it did false-positive replacements that I had to undo by hand. (This was why I started considering the "edit row buttons", so that I'd be able to skip certain types of edits manually.) To fully automate this would be a dreadful mistake. --Connel MacKenzie T C 18:08, 7 May 2006 (UTC)Reply

Why the shouting? I can read lowercase. —Vildricianus | t | 19:18, 7 May 2006 (UTC)Reply

Disagree, because this contradicts what was said above above removing templates from translation tables. I would like to see these gender/number templates used everywhere except in the translation tables.--EncycloPetey 14:31, 7 May 2006 (UTC)Reply

Um, translation tables happen to be the place where they were made for and where they're most used. In most other instances, the unabbreviated word should be mentioned. —Vildricianus | t | 15:02, 7 May 2006 (UTC)Reply

EncycloPetey, unfortunately I think you misunderstood the guideline/recommendation in the section above. This change would not be limited to translation sections, nor would it be limited to non-translation sections. This is about replacing the text ''f'' with {{f}} wherever it occurs. --Connel MacKenzie T C 18:08, 7 May 2006 (UTC)Reply

(Well, everywhere ''f'' is being used as a grammatical gender marker, that is. —scs 15:30, 5 June 2006 (UTC))Reply

Absolutely agree. —scs 15:30, 5 June 2006 (UTC)Reply

Monstrously controversial: At the bottom of a page, since numerous categories add blank lines to the bottom of a page, all categories should be listed on a single line. Note that this is a distinct fork from current practices. Effecting this massive change can only be done with semi-automatic (e.g. WP:AWB) or automatic (pywikipediabot.py) assistance. --Connel MacKenzie T C 16:54, 6 May 2006 (UTC)Reply

Hm, I'm not sure what you are saying, Connel - do you mean they should be typed on one line, or rendered by the software on one line? I was under the impression that categories do not add blank lines to a page (see #20 above). If they do, perhaps the software should be modified so that they do not. Or do you mean all on one line when they are rendered? Probably not, because that is how they appear. I don't like this proposal, in any case. — Paul G 07:21, 7 May 2006 (UTC)Reply

I'm saying that if each category is on a separate line at the bottom of the entry, the MediaWiki software make that line invisible/blank for the next pass of rendering. If there is one blank line, it is not problematic. If there are 20 categories, the intermediate processor will render 20 blank lines, resulting in a screen-full of nothing at the end of the page. Actually, I have not sufficiently tested this, so I'll do so now, on this page. --Connel MacKenzie T C 07:54, 7 May 2006 (UTC)Reply

Well, I guess I was wrong. --Connel MacKenzie T C 07:58, 7 May 2006 (UTC)Reply

Each instance of ] and all its blank spaces it produces is ignored by the software for page layout. —Vildricianus | t | 10:22, 7 May 2006 (UTC)Reply

Definitions shouldn't start immediately after the "#", but after "# " instead. --Connel MacKenzie T C 00:05, 8 May 2006 (UTC)Reply

Agree. —Vildricianus | t | 20:02, 8 May 2006 (UTC)Reply

I agree, but what about following lines with examples which start "#:" - should there be a space after the colon? — Hippietrail 20:06, 8 May 2006 (UTC)Reply

That would be #28.  :-) Please, you make a recommendation as to what's best in that case - I am fairly inconsistent with them. --Connel MacKenzie T C 20:08, 8 May 2006 (UTC)Reply

I usually put a space after #, #:, #*, *: etc, except after a single asterisk. —Vildricianus | t | 22:23, 8 May 2006 (UTC)Reply

I agree, but also, one space after a template after a # e.g. "#{{chemistry}} definition" SemperBlotto 20:51, 8 May 2006 (UTC)Reply

Yes, but I usually also put a space between # and {{chemistry}}. —Vildricianus | t | 22:23, 8 May 2006 (UTC)Reply

• Perhaps this could be better restated. "One space after line-start wikisyntax." Or: "One space between wikisyntax and visible content." The reason I am inconsistent is that "*:''" is something I confuse with "*: ''". Um, I mean without the double-quotes.
• Argh. OK, said a different way: line start wiki-syntax characters that can have a space after them, before the content starts, should. Doing so helps newcomers understand which characters are wiki-formatting and which are content-formatting. This "guideline" could apply to stackable ("*", ":", "#") and non-stackable ("{|", "||", "|}") wiki-syntax.
• --Connel MacKenzie T C 17:19, 9 May 2006 (UTC)Reply

• So is that also one space after *? If so the perfect format is imperfect. Jonathan Webley 08:17, 6 June 2006 (UTC)Reply
  • I never do that. — Vildricianus 09:57, 6 June 2006 (UTC)Reply

Thornier issues exist, like definition line "tags." For a while, we've waffled between (tag1), (tag2), (tag3,), (tag4, tag5): tag6: and many others. Recently, I was scolded for ending a list of tags with a colon, as there was already parenthesis there. But it is also improper to have a list of parenthetical items. I think the parenthesis are astronomically stupid looking now. I'd like to see that issue on the BP and voted on. Maybe parenthsis abuse is a British thing, I dunno. But we probably shouldn't have any parenthesis on any definition lines, ever.

This whole topic could also be restarted for each namespace, as each seems to have its own rules: Templates (doc on page, or in talk?,) Categories (crazy pipe partial-sorting thing,) Appendix (pseudo-namespace, Wikipedia=style heading rules), Wiktionary (1/2 Help: format, half blog format) etc.

Right now, I need sleep. More thoughts tomorrow. --Connel MacKenzie T C 08:53, 30 April 2006 (UTC)Reply

• Let me comment on your points. Please do not change their order or remove any since I'll refer to them by number:

comments moved into corresponding sections above

Definition tags are a problem. However I've been thinking a lot about lists lately and I think we can achive a lot by making them use simple # for editing them, and CSS to format them, probably with some kind of ((list)) and ((end)) templates to contain the CSS classes and id's. Our current systems cannot handle placing multiple tag templates within one set of parentheses and muliple parentheses are ugly. Another simpler idea is to just forgo the parentheses as many print dictionaries do. Italics alone should suffice. My pet peeves are: 1) mixing italicized and non-italicized parentheses. 2) italicized parentheses altogether. 3) use of colons together with parentheses.

The definition tags need perhaps a whole separate conversation. Whatever we end up with has to adhere to the KISS principle. Multiple parenthesis are more than ugly; they are incorrect. More and more, I think we should not have parenthesis at all. Your point #3 is exactly what people complained about me doing. I think multiple parenthesis are far worse, but that is my POV. --Connel MacKenzie T C 23:41, 30 April 2006 (UTC)Reply

As an interim compromise, I would propose that we do recommend using templates for definition tags, always. Rendering (whether italic or not, parenthesized or not, abbreviated or not, etc.) can be handled by the template expansion and by Hippietrail's CSS magic. The only hard problem is how to deal with multiple tags (e.g. UK-specific vulgar computer slang), but I'm willing to defer that, and to live with redundant, uncollapsed parentheses today. —scs 15:36, 5 June 2006 (UTC)Reply

On other namespaces, I think it would be too distracting to worry about them at this stage. I think more can be accomplished by concentrating on the main area first. The rest can be addressed when we see our progress. — Hippietrail 16:32, 30 April 2006 (UTC)Reply

I agree. But I'll be thinking of them as we go along. --Connel MacKenzie T C 23:41, 30 April 2006 (UTC)Reply

Speaking of "topics deferred", do we ever want to address the formatting of pronunciations, the handling of the several different schemes, etc.? —scs 18:05, 5 June 2006 (UTC)Reply

• WARNING: I plan to put this mess onto subpages. Each of the twenty plus points need separate conversations, apparently. --Connel MacKenzie T C 23:41, 30 April 2006 (UTC)Reply

• As a fervent normalizer (12k edits that rely heavily on this business), I'd like to add some things. I understand from the above points that this is more or less the desired output:

{{see|Word}}
{{wikipedia}}
==English==

===Noun===
{{infl}}

# After a space, a definition with first-letter uppercase and full stop.

====Related terms====
*]
*]

====Translations====
'''gloss'''
{{top}}
*Dutch: ] {{m}}
{{mid}}
{{bottom}}

'''other gloss'''
{{top}}
*Dutch: ] {{f}}
{{mid}}
{{bottom}}

===Reference===
{{R:Webster 1913}}

]

----

==Spanish==

===Noun===
'''headword'''

# ]

]

]

This is not entirely the way I've been doing it but I understand that it's largely what Connel's points involve. Right? Never mind the placement of categories now.

Individual comments moved to appropriate section above.

On tags: I've been thinking and looking to what print dictionaries do, and have found the SOED, which uses italics and small caps, to have a nice format. Thanks to the templates we use for them, we can change this overnight, although we'll need to instate commas where necessary. To give you an idea, it looks like this:

1. AERONAUTICS Definitions.
2. NAUTICAL, CHEMISTRY Two tags here.

It's of course also possible to "solve" the tag issue by adding the parentheses manually and leaving them out of the templates, like ({{biology}}, {{chemistry}}). —Vildricianus 18:27, 30 April 2006 (UTC)Reply

Yes, definitely. — Paul G 07:21, 7 May 2006 (UTC)Reply

Note: I added the blank line after the inflection line in your example above, but I left the "incorrect" category placement as that is so touchy. --Connel MacKenzie T C 23:41, 30 April 2006 (UTC)Reply

• NOTE: (as I said above) I plan to move all of this to subpages here. Each numbered list item above will become a heading and I'll rearrange some of the related ones after adding yours and HT's comments into each section. --Connel MacKenzie T C 23:41, 30 April 2006 (UTC)Reply

• Hrumph. Maybe I'll make them sub-pages tomorrow. --Connel MacKenzie T C 06:14, 1 May 2006 (UTC)Reply

• Yup, fine. I'd concentrate for this round on the 20ish points, leaving discussion about tags and category placement for later (the latter is "ongoing" at BP). —Vildricianus 07:59, 1 May 2006 (UTC)Reply

• I'm finding it difficult to relate the comments to the points, but in general I agree with the sample layout that Vildricianus layed out above. There are just a few points that I would disagree on or don't understand:

Individual comments moved to the appropriate section above.

Oops, I forgot to comment on this first time round. I prefer to put the {{wikipedia}} tag between the POS and the inflections. Not only does this put the Wikipedia link along side the word to which it refers, it also usually avoids adding any blank space to the rendered page. Putting it at the top or elsewhere can often do this, making the page look inelegant. — Paul G 07:27, 7 May 2006 (UTC)Reply

My convention is dependent on the number of headings in an entry. If there are more than three headings, an automatic TOC will be generated. When a TOC is present, it makes sense to have the {{wikipedia}} link above the first heading so that it appears in the white-space to the right of the TOC. But if there are less than four headings, I do exactly as Paul says: between the POS heading and the inflection line. --Connel MacKenzie T C 07:50, 7 May 2006 (UTC)Reply

Regarding terminology: "heading" vs. "header". I understand the confusion that exists with the overlapping of terms with regard to C/C++ ".h" files. However, we are in a Wiktionary context, not a C or C++ context. I use "heading" to refer to a rendered == == section heading and "header" to refer to the wikitext that denotes it. I'd prefer to not change my informal convention, as the distinction is perhaps helpful, sometimes. --Connel MacKenzie T C 13:46, 2 May 2006 (UTC)Reply

Actually there are other senses I was think of such as in binary file formats. Encarta does however give a sense for "header" that is pretty much a synonym for "heading". Collins doesn't but hey this is all just trivia. Back to the topic (-: — Hippietrail 22:55, 2 May 2006 (UTC)Reply

I think this thread is a valuable resource regarding Wiktionary terminology (things like "inflection line", "POS headings" etc.) Anyone feels like expanding Wiktionary:Glossary? Or shall I put it on my Todo? Having a comprehensive reference list of these is beneficial for mutual understanding on all this matter. —Vildricianus | t | 10:29, 7 May 2006 (UTC)Reply

Please check this edit to see that it conforms to the "perfect entry" and all rules, as above. I suggest this be the format for "phrasbook" entries. Perhaps {{phrasebookdef}} for the definition line? --Connel MacKenzie T C 16:12, 6 May 2006 (UTC)Reply

Specifically rule #23. --Connel MacKenzie T C 17:06, 6 May 2006 (UTC)Reply

I think we are nearing the wrap up time for this, as Vild pointed out in the top-most section. I would like to thank everyone who has participated. I also request that each continue to participate.

Vild, HT or I will take this list, glean the least controversial dozen or so items and build a small list to post on the BP in the next day or two. Some rewording of HT's preamble may be in order. My initial list had some pretty bone-headed stuff, that I think you've all helped clear away.

I have enjoyed (perhaps the most) Paul's linguistics lessons, even if I do still disagree. I did not expect this to be so educational, but I am glad it was.

Thank you all again. Let's see where we take this next, eh?

--Connel MacKenzie T C 18:15, 7 May 2006 (UTC)Reply

• As an aside, I just had a good laugh reviewing Talk:Main Page#Uniform data formats (give a bot a chance). I'll note that only one item in this list matches any of my complaints on Talk:Main Page. --Connel MacKenzie T C 06:17, 17 May 2006 (UTC)Reply

How about wrapping up? I could make a summary of this, put it in the Beer parlour and expect a couple of people to laugh at us. Most of the above described format rules are exercised by my javascript, though, so given enough time, all entries will move towards this standard (especially the blank lines stuff, which is largely at odds with what most entries have and most contributors do). — Vildricianus 11:57, 5 June 2006 (UTC)Reply

You guys may be ready to "wrap this us", but remember, some of us just got here. :-) Me, besides the comments I just sprinkled above, I've got one biggie: some of this stuff matters much more (or at least, differently) than others.

There are several quite different goals someone might have in pursuing this normalization:

(A) Make pages look consistent to readers.

(B) Make page source look consistent to editors.

(C) Enable automatic processing, today, e.g. by using templates for gender and definition tags.

(D) Enable machine parsing of wiktionary entries (for other purposes).

(E) Turn Wiktionary into the more-structured database some of us wish it were.

Of these, to me, (B) is the least important, although about half of the entries on the normalization list have to do only with this. A reader doesn't care about blank lines before or after headers, or extra spaces between the == == and the heading text. And a (reasonable) parser doesn't care, either. Obsessive editors might care (and I don't mean to be critical when I say "obsessive" here, as there are plenty of these things that I'm obsessive about, too), but to me, if an entry is complicated and hard to edit I'll add a few blank lines here and there without thinking about it too much, and if an entry is too sparse I'll delete some blank lines without thinking about it too much, and I really don't worry about "standardizing" this sort of thing. A guideline for the preferred formatting at this level would certainly be useful, but it doesn't have the same kind of urgency as the other four goals I've listed.

Of these, to me, (D) and (E) are quite important, although I know that there are lots of editors for whom they're not important at all, who are just as uninterested in machine parsing as I am in normalizing newlines. And there's nothing wrong with that, either; there are legitimately multiple goals here (only partly overlapping, or sometimes wholly distinct). (So what I'm saying here is that it doesn't bother me if someone is uninterested in machine parsing, or interested in blank lines, as long as they're not bothered by the fact that I'm not.)

But I bring this up because I think the larger audience is going to be equally struck by what they'll see as the same kind of differential importance of these issues. So I think we need to at the very least split them up, and perhaps even address them as somewhat separate subprojects. Here's my take:

issues that matter for proper functioning of existing software

3, 4, 5

issues that are visible to readers

7, 8, 18, 23, 24

issues that enable wiktionary-specific processing, visible to readers

25, (definition tags templateized)

issues that are visible only to editors (cosmetic)

1, 2, 6, 10, 11, 12, 13, 14, 15, 16, 19, 27

issues that are visible only to editors (possibly more significant)

20, 21, 22

issues that matter for machine parsing

7, 8, 20, 23, 24, 25, (definition tags templateized)

not sure (including stricken)

9, 17, 26

There's some overlap there; in particular, everything I've listed as "issues that matter for machine parsing" is also in some other category (usually, interestingly enough, "issues that are visible to readers").

I have included "definition tags templateized" (which was in Connel's "Topics deferred" section), because I think it's as least as interesting as some of the other issues listed here, but not as controversial as some. Call it #28?

—scs 16:30, 5 June 2006 (UTC)Reply

1/ I have a gut feeling that there are more editors than readers. 2/ This list was initially intended for the hardcore normalizers like Connel, Hippietrail and a few others. Other options got stuffed in afterwards and it turned out to become a more general discussion. So (B) was really the first purpose of this thread. Having the blank lines appearing correctly is not very important, but fun for those who care and easy for a javascript to add. It also allows for quickly spotting "bad" entries, i.e. those with everything compressed, which is usually done by the less experienced here, and which are therefore possibly "bad". 3/ The other options you presented are also worked on, for instance with the inflection templates. I'm not sure though how urgent they are. Machine-parsing Wiktionary has only been done by the few (Connel, Patrik,...) who wanted to put up todo lists or stats. There's much more work on the actual content before we should consider any other fancy option. That doesn't defer normalizing for machine parsing, though. But it's not "urgent". It's work in progress just as much as the other jobs are. Compare the format of two years ago with the current. Also keep in mind that it changes constantly. 4/ I'm really bothered by the fact that you don't like the blank lines :-(. 5/ Thanks for commenting on it. I guess it's a topic that will pop up from time to time in order to get one another's confirmation of existing practice (which is what I think Hippietrail intended to do). — Vildricianus 17:47, 5 June 2006 (UTC)Reply

Agreed that machine parsing is a longer term work-in-progress -- but note that it applies even though (nay, because) the format "changes constantly": if/when we do a big format change, the more we can automate it, the happier we'll be. (Now, as for those cotton-picken blank lines, I don't just "not like" them, I can't stand them, I fly into a blind rage when I see them, I delete them on sight, I'm really bothered that you could consider sneaking your slovenly blanklineist POV into this vital emerging policy...) —scs 18:27, 5 June 2006 (UTC)Reply

• I wish to say that I support any coherent, logical subdivision of this monster (that I thought at first was so short and sweet) subpage. --Connel MacKenzie T C 18:20, 5 June 2006 (UTC)Reply