Wiktionary talk:About Ancient Greek/Declension table cleanup

Declension template cleanup
It seems to me that is redundant given. Is this the case? I anticipate there may be several others that are redundant. More generally, there seems to be a fair number of templates named against the general  format. Ought we get a bot and correct these? Indeed, many of the templates do not use the template, do not have the correct titles and notes formatting, and do not have the correct macra and breves. —JohnC5 06:59, 12 February 2015 (UTC)
 * Gah, my bad, I derped. And yeah, there are a lot of bad templates, although I think it may be better to drop in favor of simply using, that is to say,
 * which, if I'm not mistaken, should function exactly like, except without the unnecessary (and deprecated) checking-for-page-existence.
 * And yes, someone needs to fix those template names. I'll get on it. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 16:01, 12 February 2015 (UTC)
 * So at this point the only ones that need work are:
 * Also, where did these declensions come from? I believe that these diaeresis declensions exist, but I would love an example of a word that uses them. Also, could you look at the length of the upsilon in the dative plural of ? Should it be long or short, given the same case in ? —JohnC5 23:36, 12 February 2015 (UTC)
 * I saw the changes you made. Derp. —JohnC5 23:37, 12 February 2015 (UTC)
 * Re, it's entirely true that we should be showing alternative forms, but first I'd like to have dialectical conjugation templates... which I'd like to get around to creating soon, actually. In the meantime I'd just put down the exceptional ones... or help create the dialectical templates, if you can. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 01:11, 13 February 2015 (UTC)
 * I've emptied out the trasclusions for so it may be removed. Relatedly, I cannot figure out why  is not drawing its categories from its documentation page (at least at present). Are the servers just taking their sweet time to update? Please forgive my noobishness.
 * I would love to attempt to help create new dialectal templates with your aid, advice, and full knowledge that I have never taken any AG classes in my life. I can research with the best of them though.
 * Before we go adding a bunch of new templates, it seems to me we should sort out this / issue for the existing templates and establish a template nomenclature system for dialectal templates. It bothers me like crazy that is not / or something of the ilk. Might I suggest we use the general principle of either
 * or
 * and using the codes defined in Module:grc:Dialects?
 * Also, thanks so much for your attention and help with my weirdo requests! —JohnC5 08:32, 13 February 2015 (UTC)
 * Also, and  may now be removed. :) —JohnC5 10:47, 13 February 2015 (UTC)
 * Yes, the servers take a while to update those. And I'd think 2nd-att is better... I'd been going for something like grc-decl- - - - -, putting dialect near the end because most types will end up having slight variations for each of the dialects. I'll go and delete those orphaned templates, thanks for helping. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 14:09, 13 February 2015 (UTC)
 * More to be removed:
 * —JohnC5 17:48, 13 February 2015 (UTC)
 * Also, I'm not sure who is in charge of, but it hasn't been working for a few days now... Might I suggest we switch over Bible Hub's or Blue Letter Bible's implementation of Strong's? —JohnC5 18:43, 13 February 2015 (UTC)
 * Re, I'm wondering if it might be better to shift the calling of to an intermediary grc-cell template, so you could call it with  instead, which looks smaller to me (but I know nothing about optimization...) ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 22:15, 13 February 2015 (UTC)
 * I had the same thought, actually! Maybe we could get a second opinion? The current already uses ; though, I don't want to change it at present because it would wreak havoc until we changed everything over. Maybe I could set up, implement it, add it to all the entries then change them all back with a bot in one fell swoop. Or we could come up with a better name entirely (I dislike grc-cell for no obvious reason) and just change everything over? —JohnC5 23:08, 13 February 2015 (UTC)
 * I've also created based on §704 of Pharr and the LSJ entries for  and . Is this to your liking (excusing the  business to be fixed)? —JohnC5 04:51, 14 February 2015 (UTC)
 * And can be removed! —JohnC5 05:36, 14 February 2015 (UTC)
 * and may be removed. —JohnC5 04:59, 15 February 2015 (UTC)
 * Also, and sorry to keep pestering you, I believe and  are probably misnamed, right? —JohnC5 00:49, 16 February 2015 (UTC)
 * Yeah, I think the homeric template looks good, and yes those are misnamed and I'll fix them. I hadn't bothered to look at adjectives yet. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 03:57, 16 February 2015 (UTC)
 * may be removed.
 * Shall I begin making Homeric templates from Pharr? And should we create subcategories like Category:Homeric Greek noun inflection-table templates for the sake of maximum unnecessary categorization? —JohnC5 08:17, 16 February 2015 (UTC)
 * Do that, but I'd recommend cross-checking with Smyth and LSJ as well. And let's leave off the categories for now. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 15:31, 17 February 2015 (UTC)
 * may be removed. Could you check and ? —JohnC5 21:20, 19 February 2015 (UTC)
 * Howdy! I know there was a bit of a pause. I was wondering whether we should implement the replacement then go through all the templates and update them to the present standard:
 * title, titleplural, and titleapp
 * grc-cell
 * notes (and notes1, notes2, etc. for templates with multiple tables)
 * correct vowel length
 * Is there anything else that should be added to the templates? It seems like we could edit the overall template structure to have it automatically add the title given just the declension number and dialect, and have the lower level templates just draw from the cells in the body of the table. Does that make sense? I will start as soon as we have the new form of the grc-cell. Should we bring someone in to write it for us? I just realized we hadn't worked on this for a bit. —JohnC5 09:47, 28 February 2015 (UTC)
 * and may be removed. —JohnC5 04:59, 15 February 2015 (UTC)
 * Also, and sorry to keep pestering you, I believe and  are probably misnamed, right? —JohnC5 00:49, 16 February 2015 (UTC)
 * Yeah, I think the homeric template looks good, and yes those are misnamed and I'll fix them. I hadn't bothered to look at adjectives yet. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 03:57, 16 February 2015 (UTC)
 * may be removed.
 * Shall I begin making Homeric templates from Pharr? And should we create subcategories like Category:Homeric Greek noun inflection-table templates for the sake of maximum unnecessary categorization? —JohnC5 08:17, 16 February 2015 (UTC)
 * Do that, but I'd recommend cross-checking with Smyth and LSJ as well. And let's leave off the categories for now. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 15:31, 17 February 2015 (UTC)
 * may be removed. Could you check and ? —JohnC5 21:20, 19 February 2015 (UTC)
 * Howdy! I know there was a bit of a pause. I was wondering whether we should implement the replacement then go through all the templates and update them to the present standard:
 * title, titleplural, and titleapp
 * grc-cell
 * notes (and notes1, notes2, etc. for templates with multiple tables)
 * correct vowel length
 * Is there anything else that should be added to the templates? It seems like we could edit the overall template structure to have it automatically add the title given just the declension number and dialect, and have the lower level templates just draw from the cells in the body of the table. Does that make sense? I will start as soon as we have the new form of the grc-cell. Should we bring someone in to write it for us? I just realized we hadn't worked on this for a bit. —JohnC5 09:47, 28 February 2015 (UTC)

Ah, sorry, I had been away. I think the Homeric templates look fine, although I'm surprised Smyth doesn't mention -ῃιν if it's a dual ending. That looks about right as far as I know. Re grc-cell, it really seems like it's probably better just to do away with the 1v/NSv arguments; they're a relic of the time before automatic diacritic removal, and I'm struggling to think of a way in which they'd still be useful. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 21:51, 1 March 2015 (UTC)
 * Pharr mentions both and that scholars read it differently (§654, 1st Ed.), though my copy of Pharr is old and may have be updated. Smyth only mentions one.
 * I concur. I wonder how we can find which AG entries use the 1v/NSv arguments now, so we can switch them over and retain the information once we have changed over to the new cell format. Bots? So this means that effectively the new version would be and would represent something like ? —JohnC5 22:25, 1 March 2015 (UTC)
 * I went ahead and made . Tell me if it looks correct. Also, feel free to change/move/delete it, etc. —JohnC5 06:20, 2 March 2015 (UTC)
 * Actually, given that we don't need tr= if we're using l-self (I was unaware that template existed, even), and that the question mark seems superfluous, I imagine we could probably do away with any sort of cell template. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 20:50, 2 March 2015 (UTC)
 * the template documentation claims that the  does not do anything, but a test like this shows it is false:
 * Without :
 * With : ‎
 * And it does straighten out the input, which is pleasing. So shall we keep it? —JohnC5 21:16, 2 March 2015 (UTC)
 * Then tr= shall be kept (and I'll go fix the documentation for ) but even then I suspect it's better to leave off using a cell template. (Alternatively, I'll bring up in the beer parlour whether we want to remove transliteration from l-self anyway.) ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 12:56, 3 March 2015 (UTC)
 * Fine :/. I was also thinking that I'm going to change the  field in all the grc-decl-blank-* templates so that it reads:
 * This way, all the templates will at least read "Declension of x, y." Then we can do away with the hideous  and   arguments, and replace them with:
 * It has always annoyed me that we have to specify NS, GS, NP, GP, titleapp, and the declension twice in every template and it looks very ugly. What do you think? —JohnC5 18:52, 3 March 2015 (UTC)
 * Sure, why not. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 19:05, 3 March 2015 (UTC)
 * Well that took slightly longer and more fiddling than anticipated, but now it is done. Changes:
 * no longer exists
 * overrides the title completely (for use in entries)
 * is used in templates to add to specifiers (dialects, contraction, etc.)
 * is used in entries to append to the title.
 * is used in templates to specify the declension information
 * The titles should now draw their forms from the appropriate nominative and genitive cells.
 * In the process, I thought I would need these templates (,, and ), but I was incorrect. Please remove them (though do not remove , as it is still important). All that is left to do is to decide on the cell format, then implement it and check that all the vowel lengths are correct. I hope the new format of the templates is not too jarring to look at. —JohnC5 08:41, 4 March 2015 (UTC)
 * Deleted, and many thanks. I'll start fiddling with the API and see if I can't figure out a way to check for 1v/NSv arguments. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 14:27, 4 March 2015 (UTC)
 * I've noticed it's common practice to wrap the notes sections of templates in templates. Should I add this into the templates directly, or should we leave this for the entry creator? FYI, I'm trying to standardize the templates such that two-table templates have ,  ,   arguments for both tables, the first table, and the second table respectively. —JohnC5 22:43, 4 March 2015 (UTC)
 * is used in templates to specify the declension information
 * The titles should now draw their forms from the appropriate nominative and genitive cells.
 * In the process, I thought I would need these templates (,, and ), but I was incorrect. Please remove them (though do not remove , as it is still important). All that is left to do is to decide on the cell format, then implement it and check that all the vowel lengths are correct. I hope the new format of the templates is not too jarring to look at. —JohnC5 08:41, 4 March 2015 (UTC)
 * Deleted, and many thanks. I'll start fiddling with the API and see if I can't figure out a way to check for 1v/NSv arguments. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 14:27, 4 March 2015 (UTC)
 * I've noticed it's common practice to wrap the notes sections of templates in templates. Should I add this into the templates directly, or should we leave this for the entry creator? FYI, I'm trying to standardize the templates such that two-table templates have ,  ,   arguments for both tables, the first table, and the second table respectively. —JohnC5 22:43, 4 March 2015 (UTC)

Definitely add it directly, and use as it's shorter. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 22:51, 4 March 2015 (UTC)
 * I've added it to the grc-decl-blank-* templates so you won't have to deal with it in any of the higher level templates. —JohnC5 23:12, 4 March 2015 (UTC)
 * Could you look over and check that is represents the most up-to-date format? —JohnC5 09:37, 5 March 2015 (UTC)
 * Actually, I've updated again so that it will do all the work that could need to happen in a cell, including when there are multiple forms in a cell. See if you find that acceptable? —JohnC5 23:59, 5 March 2015 (UTC)
 * poke! — JohnC5 05:05, 16 March 2015 (UTC)
 * Whoop, I'd forgotten. The Homeric template looks right, although I'm surprised Smyth doesn't notice -ῃιν. And sure, we may as well keep a cell template. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 13:49, 16 March 2015 (UTC)
 * Cool, then I shall begin the process of converting them all to the this format and making sure the vowel lengths are correct. Should we add documentation to them all? Also, I think -ala and -alp are reversed in the table on Category:Ancient_Greek_1st_declension_templates, perhaps? — JohnC5 19:35, 16 March 2015 (UTC)
 * Also, given that Μανασσῆς should have genitive and accusative in -ῆ according to Strong's and that it is the only transclusion of Template:grc-decl-1st-M-eta-con-η-gen, perhaps this template ought be removed? I think an overridden version of Template:grc-decl-1st-M-eta-con would work fine for Μανασσῆς. — JohnC5 20:02, 16 March 2015 (UTC)
 * None of the transclusions for grc-decl-1st-ala-pax are actually -pax. They are either disyllabic -prx or . Thus, I think should be removed
 * Ok, I've removed all the transcludes. Please kill it. — JohnC5 23:42, 16 March 2015 (UTC)
 * Done and done. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 01:54, 17 March 2015 (UTC)
 * I've also removed all the transcludes for because they were all just disyllables in the same situation as above. Please kill this also? It's fun removing bad templates! :) — JohnC5 02:23, 17 March 2015 (UTC)
 * Also... should be removed. In fact no template should have both -pax and -prx (at least in the first declension). — JohnC5 02:51, 17 March 2015 (UTC)
 * Actually, I'm not convinced that all the entries in and  should not be entirely supplanted by . Shall I change them all over? These seem only like dialectal differences. — JohnC5 03:20, 17 March 2015 (UTC)
 * Sorry to keep pestering. I have move the content of into, so could you delete the former? When this is done the AG templates should be fewer in number and more sensible. — JohnC5 03:56, 17 March 2015 (UTC)
 * Actually, hold off on deleting, , and (but still delete  and ). I'm in a quandary. For an entry like θάλασσα or γλῶσσα where we have multiple dialectical stems which take different endings, do I
 * Make a unified template which allows for specifying different stems for each dialect but which otherwise auto-generates all the dialects using the same stem?
 * Leave these variety of separate templates?
 * The problem with option 1 is that the template becomes more complicated to use and would have different stem nouns inside one entry (not that you wouldn't still create the other entries for each dialectal stem). On the other hand, option 2 has a variety of confusingly overlapping templates that do not offer the completeness of a unified template. If I went with option 1, I would make the other 3 templates back-end templates to be used by grc-decl-1st-prx and not separately in entries. With option 2, I'm not sure how we would fix the problem. Sorry to leave these contradictory requests. Also, at some point should we start a new thread (This one is getting a bit long)? — JohnC5 06:40, 17 March 2015 (UTC)
 * I created for words in -ᾰ that had attested genitives in -ᾱς and -ης—that is, it had to be attested in both dialect groups. I don't think the Homeric templates should be folded in to that, and I'm certain that the simple first declension templates shouldn't be. On the other hand, I have to wonder if there are few enough dialectical differences in general (except perhaps Homeric) to justify just making cells like  . Or maybe we should leave off putting dialectical forms in the templates in favor of putting it all in Appendix:Ancient_Greek_first_declension. I honestly don't know which is the best option. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 13:44, 17 March 2015 (UTC)
 * So as it currently stands, the transclusions of grc-decl-1st-ets-prx and grc-decl-1st-ala-prx are divided exactly along "party lines," which is to say, all the -ala transcludes have stems in -i-, -e-, and -r- and -ets has all the rest. This seems to be a pretty good indication that they are divided according to the Attic sound change. That indicates to me that we could fold them into one template.
 * I'm fine with not folding them in, but we would need to make the use of more clear than it is. I'd be fine adding the dialectic endings into one template or creating a whole slew of separate dialectical templates then folding grc-decl-1st-ets-prx and grc-decl-1st-ala-prx into one Attic template. Actually, there seem to be several different options, but I'm not sure I like it as it is. Regardless, I will fix grc-decl-1st-prx however we decide. — JohnC5 21:35, 17 March 2015 (UTC)

There are two reasons to not combine ala and ets, that is to say, two reasons why a page might use one directly: (1) words that are only attested in one dialect, and (2) words like γλῶσσα that have different lemmata for one or more dialects anyway. That's not to say we couldn't put switches in a combined template to account for those, though-- I'd imagine it would be relatively simple. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 22:46, 17 March 2015 (UTC)
 * So, to make sure I got this right: The template would need to:
 * Allow for specification of stems by dialect
 * Allow for specification of dialects of be shown (alternatively, specify which dialects should not be shown)
 * Is that right? Could we do something like . I'll also point out that I can use to detect whether a stem ends in ι/ῐ/ϊ/ῑ, ε, or ρ and sort it into its correct Attic declension. Is this what you had in mind? Also, do you know of a template that spits out the form of the word by created by autolinking (i.e. that would take ἰᾰσπῐ́δοιν and make ἰασπίδοιν)? — JohnC5 03:09, 18 March 2015 (UTC)
 * What you're looking for is makeEntryName in Module:languages, but that isn't accessible through template alone. I have to wonder if it wouldn't be better to Luaïze the declension tables anyway. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 14:40, 18 March 2015 (UTC)
 * After all this work? You are cruel. :) We could do that, I guess. I'd be prepared to work on that with you as I am less familiar with the utilities of Lua.
 * I'm surprised there's not such a template like . It seems like a useful and easy utility to program.
 * Also, take a look at which I'm using to generate documentation for these pages like.
 * Also, I'm not sure why the override parameters (NS=, etc.) need to replace the autogenerated form. In most cases, it seems like they should be adding and not replacing. — JohnC5 18:39, 18 March 2015 (UTC)
 * Whoah, Nelly, I didn't say you had to do it. I'm quite fine with writing it myself if that's what we decide. :-)
 * I have to ask, though—why do we need a template to strip length marks? What were you planning to write? ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 21:03, 18 March 2015 (UTC)
 * Oh I was thinking of automating the  feature in templates like  (or the genitives in ), which can be done with . Length marks make it slightly more complicated (indeed, I could still do it with some slightly more complex logic), and I thought such a template might be nice to make it easier. — JohnC5 21:15, 18 March 2015 (UTC)
 * I'm not sure str rightmost works anyway? (trying it normally yields ) although I may be using it wrong. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 02:32, 20 March 2015 (UTC)

You have to specify the number of characters to show from the end. So: Why an offset of 0 doesn't return nothing is beyond me. — JohnC5 03:47, 20 March 2015 (UTC)
 * Then I don't know. If it's better to write a module, I'll do that, otherwise, we'll create and leave declension to the templates. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 14:25, 20 March 2015 (UTC)
 * For the sake of argument, here are the pros and cons of each (from my perspective):
 * Module&#58;
 * Pros&#58;
 * Allows for greater string manipulation (for alternative forms and dialectal endings, etc.)
 * Could potentially automate the placement of accents1
 * Perhaps more resource efficient than templates? (I have no idea)
 * Eliminates the need for  backend
 * Cons&#58;
 * Less accessible to a normal user who wants to understand how the template works
 * Harder to add new templates (dialectal, etc.) for normal users
 * JohnC5 will probably have to pester ObsequiousNewt more often :)
 * Template&#58;
 * Pros&#58;
 * They are comparatively easy to use and understand from the code
 * They don't currently suck
 * JohnC5 and others have wasted a good deal of time on them (Pro?)
 * Cons&#58;
 * There might be less efficient than a module (again, I don't know)
 * There are a few string manipulation techniques that are harder to achieve
 * Notes&#58;
 * 1I believe I have run across compound words where the second element is disyllabic that do not permit a recessive accent to travel to the first element of the compound. It seems like automatic accents would have trouble accounting for this, but this could simply be overridden by user input. On the other hand, automating the accents is probably not strictly necessary.
 * Have I missed anything? — JohnC5 20:59, 20 March 2015 (UTC)
 * That certainly seems to cover most of the arguments, and based on that alone, I'd lean in favor of templates. On the other hand, we also need to decide the best place to put dialectical forms. Here are the options I see:
 * In the same cell as the normal form (may cause cells to be too large, and also doesn't allow for as much in terms of more detailed notes)
 * In a separate table (can mean a lot of tables when the inflections may only vary by one or two forms)
 * In footnotes (may mean a lot of footnotes)
 * On a separate page like Appendix:Ancient Greek first declension (not very visible)
 * Honestly, I'm leaning towards the footnotes myself, since having a lot of footnotes (although, frankly, I don't know just how many it would end up being anyway) doesn't seem too bad. Thoughts? Other options? ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 19:34, 22 March 2015 (UTC)
 * I think footnotes would be most appropriate, but I think I will try make a separate backend that will allow things like  That will automatically add reference numbering to the table. Alternatively, each dialect could have it's own set (,  ,  , ...). What do you think? — JohnC5 23:45, 22 March 2015 (UTC)
 * Customization is going to be difficult. I'd lean towards the first option, as there are about two dozen dialects and it'd be easier to write. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 01:53, 23 March 2015 (UTC)
 * Ok, I'll begin adding them. Should there be a way to turn them off? I think I'm fine with leaving things like Homeric forms as long as we mention not all forms exist for every word in every dialect. — JohnC5 03:00, 23 March 2015 (UTC)
 * Ok, check out . It's only implemented via the, but I can expand it. I tried to use ref tags, but they were not working. If you can figure out why, I'd far prefer ref tags. — JohnC5 04:12, 23 March 2015 (UTC)
 * I feel like I'd make things a little less conspicuous. Like, I like the way does its footnote. I don't know why the ref tags don't work; if you can put a broken version in a sandbox I'll try and fix it. We also may want a way to automatically replace the cell proper with a dialectical form in case the stem varies and we can't just do a footnote (like, say, Doric δᾶμος.)
 * I've been working, for the past week or so, through Smyth and Buck, taking notes on all the dialectical forms. (Doric, by the way, is a goat rope, and I often have no idea whether or how to simplify it.) These come down to (1) different forms, like Lesbian -μεναι for -ναι, (2) different stems or whole tenses, like Doric -έω in the future, and (3) differences (in either forms or stems) caused by phonology, like Attic-Ionic η for ᾱ or Arcadocypriot -υ for -ο. To which I say (1) footnotes, probably whether a dialectical form is attested or not—but this is debatable, (2) another conjugation table, probably iff the dialectical form is attested, and (3) I'm not really sure when, or even whether, this should be shown. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 00:24, 28 March 2015 (UTC)

Ok, I've made User:JohnC5/Sandbox1, User:JohnC5/Sandbox2, User:JohnC5/Sandbox3 which contain altered versions of, , and a test area respectively. The problem seems to be that, while the tags will appear in the correct places, their content will not, even if it is just plain text. I can't figure out why. Also, feel free to fiddle around with the style, etc. I will try to think of the best way to replace the forms within the table. This would also help for the true Homeric declensions as well as the Doric ones as you mentioned. — JohnC5 04:41, 28 March 2015 (UTC) (Some hours later) I've fixed the ref's to work correctly. Please check out the styling, etc., and see whether they are to your liking. — JohnC5 06:48, 28 March 2015 (UTC)
 * Well, I switched it from to a &lt;references /&gt;, because reflist had weird formatting issues. Quite frankly, if it were up to me, I'd get rid of using references (the only benefit we really get is automatic numbering) and just put a 1 etc. tag after the form, because my eye wants it to look as inconspicuous as possible, and even [note 1] seems too large. But that's the only real reason I have for changing it. Meanwhile, do you have opinions on the three types I mentioned above? ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 19:36, 28 March 2015 (UTC)
 * I like the idea of footnotes forms that have the same lemma form. This should also be led informed by what other dialects are known to exist for a word. If another lemma form is attested, we make a full table conjugation on another entry. Does that make sense?
 * As to the question of reference numbering, I worked so hard on this because I was annoyed by the concept of the numbers starting at 3 at the first reference in the table. Let me check to see if I can't use variables to replicate the effect of references without the linking, etc. I also was think of conditionally adding a double border before dialectal forms or notes just to offset them. Let me know how that looks. — JohnC5 20:28, 28 March 2015 (UTC)
 * Never mind. It turns out the var extension is not installed. :/
 * Ok, I've updated it. I left in the 3 column version, but if you don't like it, just remove the and . — JohnC5 21:25, 28 March 2015 (UTC)
 * Ah, yes, I'd forgotten why you needed to use footnotes in the first place. Well, let's make it [1] at least (it's not like inline citations are really a thing in Wiktionary.) The columns also seem odd to me. But at this point, neither thing really matters, so long as the information is there. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 04:44, 29 March 2015 (UTC)
 * Take a look at the sandboxes now, and tell me what you think. I've also gotten the versions of the base templates as close as possible. Also, I've rolled the functionality of into, so grc-decl-blank-plur0 may be deleted. — JohnC5 05:01, 29 March 2015 (UTC)
 * No, no, put the ref tags back. I was complaining for no reason and now the numbers are out of order. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 13:47, 30 March 2015 (UTC)
 * Ok, I've changed them back, though I used a as the "group" just to ensure that there are no accidental reference inclusions (I personally tend to use references in etymologies from time to time). There is still a leading space on the citation superscript, but I think you'll only notice if you look for it. The columns, however, are up to you―take them or leave them. — JohnC5 04:06, 31 March 2015 (UTC)
 * As for exchanging the forms in the table itself with dialectal forms, I was thinking about this:
 * Have which is effectively a switch statement which defaults to Attic. This would only be used on a table cell that would have more than the normal Attic forms (otherwise just a grc-c would suffice).
 * To switch the table to a dialect, you type dial=hom/dor/ion/...
 * to add footnotes for a dialect you do hom=1, dor=1, ion=1, ...
 * Does what I said make any sense? — JohnC5 04:06, 31 March 2015 (UTC)
 * Yeh, I don't see any problems with that design. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 19:43, 31 March 2015 (UTC)
 * Although, if you feel it'll end up being too difficult or tedious to code, I'm willing to write a module. Just so we know that option's still on the table. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 19:49, 31 March 2015 (UTC)
 * Nah, will be pretty straight forward. I'll do it soon. Then we just need to start coming up with all the dialectal forms. :) Do the grc-decl-blank-* templates look good enough for me to update them? — JohnC5 20:42, 31 March 2015 (UTC)
 * Quite frankly, as long as they work, I'm going to leave off complaining (any more) about æsthetics. After all, we can change them later if we so choose. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 03:08, 1 April 2015 (UTC)
 * So... after having thought about it, it seems to implement the table and footnote forms will require them being specified twice in each template which seems wasteful to me. The question then would be: do we care or do you want to Luaïze it to avoid redundancy (and to have all the other fancy benefits of Lua)? It saddens me how much work has already been done, but on the other hand, Lua would probably happen eventually anyway. I'd also love to help in any way, if you want it. — JohnC5 03:51, 1 April 2015 (UTC)

Alright, I'll get on it, soon as I can. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 15:49, 2 April 2015 (UTC)
 * What have I forgotten? :)</tt> — JohnC5 19:14, 2 April 2015 (UTC)

I'm starting a new subsection
(I've started this new section because I tire of scrolling.)
 * So here's what I've come up with. We have one template,, which does all the formatting. Individual grc-decl-whatever templates are probably best put into Lua form like , but we don't have to. Should a user wish to override a section, they put in something like GS=ου/οιο::hom/οι::the/ου:ουν::kup, which will be formatted by grc-decl (and is the same way the data is stored there.) This would output something like:
 * ; (Homeric);  (Thessalian); / (Cyprian)
 * In theory, anyway—except that the page using can specify either a list of dialects to show, or a list of dialects not to show. I don't know how best to put that into input (what do you recommend? "dial=dor/aio" seems good, but I don't know how to figure excluding dialects into that.). If the former, every dialectical form is put into the cell proper (there's unlikely to be more than like two anyway) and, otherwise, the default form (i.e., Attic) is put into the cell proper and every form (that's not excluded) is put into a footnote.
 * That is to say: the above, with no dialects specified or excluded, would put ἵππου in the cell, and every other form (formatted like above) in the footnotes. If "hom" was specified, it'd be ἵπποιο in the cell and nothing in the footnotes. If "att" or "aio" was specified, it'd be ἵππου in the cell, etc. If "att/the" was specified, it'd be ἵππου (Attic), ἵπποι (Thessalian) in the cell. If "kup" was excluded, it'd be ἵππου in the cell, and ἵπποιο (Homeric), ἵπποι (Thessalian) in the footnotes. If "att" was excluded, it'd still be ἵππου in the cell by virtue of it being the default form (and everything else in the footnotes.) (Sorry if this is overlong, but I know I can be bad at communication and I wanted to make sure I was communicating clearly.) ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 21:54, 3 April 2015 (UTC)
 * I believe I understand what you are explaining. Also, which Homeric forms should we use, Smyth or Pharr? Further, should we be using Epic or Homeric throughout, or is there a difference that I've missed? — JohnC5 19:18, 5 April 2015 (UTC)
 * I guess whichever is more recent. I'm still surprised they mention different forms. Homeric essentially means "in Homer", whereas Epic is the style inspired by Homer that was used in people like Hesiod. I don't know what actual differences exist between the two. How should dialect information be formatted? ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 13:50, 6 April 2015 (UTC)
 * So I have an old version of Pharr (1st edition, 1920), but a newer version (4th edition, 2012) exists, but I do not (currently) have access to it. Maybe I could find it in a library nearby. I think the main difference I've noticed is the genitive dual business in the first declension (-ῇῐν) (though even Pharr mentions that is could be read as -αῖ(ῐ)ν). The rest just tend to be extra forms that sometimes show up (like the genitives in Ἀΐδης).
 * If that is true about Homeric vs. Epic (that is approximatley what I guessed), then I think we should just choose one as the standard and alias the other input to it.
 * I was imagining the general form  where dialects are separated by commas and Attic is listed in cells without a dialect marker but with one in the footnotes. — JohnC5 21:33, 6 April 2015 (UTC)
 * I meant in the input, but I suppose I'll just make it an initial ~ or ! or something. Meanwhile, I lost the entire draft I had been working on, so this'll be taking a bit longer than I had planned. ObsequiousNewt (ἔβαζα|ἐτλέλεσα) 02:34, 7 April 2015 (UTC)
 * So. Ugh. The more I worked, the more I saw that trying to write a module would be a really kludgy way of going about things. So I backed off for a minute to see what I wanted things to look like, and I came up with this. Which is, well, okay, but not great. For one thing, if there were a Lesbian form it would be ὤραισῐ(ν) without the spiritus asper (and so with Cretan). And if I made a separate table for ὤρα... well, it'd be difficult. The point being that I can write a module to do all that, but there are so many little things that need customization—which makes it difficult not only to write a module but also to use it—that I really question if it's worth the effort.
 * Now, if I had my druthers, I'd just put everything in an appendix. The better option is probably to compromise, and simplify to Attic, Epic, Ionic, Doric, Aeolic, lumping Homeric in with Epic and ignoring Arcadocypriot and the Northwest dialects because they're not represented in literature. Then we either make templates like, or alternatively put it in a module, which would still be annoying, but definitely easier to write than it would be with every dialect's peculiarities represented, and then those peculiarities can go in the appendix.
 * So, those are my thoughts. If you have any comment, please share it, otherwise I'm going to keep working on the module. ObsequiousNewt (εἴρηκα|πεποίηκα) 22:25, 12 April 2015 (UTC)
 * If that is the case, would you like me to continue adding to the existing modules but only allow for those 5 dialects? If we split the full dialectal tables into separate templates, then this becomes a lot easier. What do we think? — JohnC5 04:38, 13 April 2015 (UTC)
 * Do you know what, I've been thinking about this wrong. I've got a brilliant idea which will make the module a lot easier to write. ObsequiousNewt (εἴρηκα|πεποίηκα) 17:11, 13 April 2015 (UTC)

I rest assured that, in the fullness of time, one or both of us will create a working set of templates or go mad in the attempt. — JohnC5 19:39, 13 April 2015 (UTC)
 * I've managed to jury-rig something. Just don't ask how I did it. Also, you might want to stay out of the cargo hold for a while. I'll start adding the rest of the data and put it into the mainspace Soon™. ObsequiousNewt (εἴρηκα|πεποίηκα) 21:29, 25 April 2015 (UTC)
 * &#1645;&#1645;&#1645; Shiny! &#1645;&#1645;&#1645;
 * — JohnC5 22:21, 25 April 2015 (UTC)
 * Alright, I fixed over the second declension. Check this if you will. ObsequiousNewt (εἴρηκα|πεποίηκα) 20:48, 28 April 2015 (UTC)
 * <tt>:P</tt>
 * The Attic and Epic forms all look correct. For the other dialects, I'd have to break out Smyth, which you've already done. So it looks good on that front.
 * Just a friendly reminder: the other features you need to implement include:
 * The articles
 * The notes section
 * (perhaps) Fixed column widths so that you don't get wider columns all over the place.
 * (perhaps) Thinner tables for sing, plur, and sp tables. User:JohnC5/Sandbox1 has all of those numbers, if you want them.
 * When you mentioned a while ago using semicolon as the separator for the nom and gen, I didn't understand about what you were talking. Could we think of a different symbol (tilde, hedera, smiley, something) because my eye often misses the semicolon among commas?
 * This is looking great thus far. — JohnC5 19:12, 30 April 2015 (UTC)
 * I'll change the comma to a slash (like you had it originally... yeah.) And fix everything else. ObsequiousNewt (εἴρηκα|πεποίηκα) 15:40, 3 May 2015 (UTC)
 * kthxbai — JohnC5 23:54, 3 May 2015 (UTC)
 * Consonant stems, excluding σ, are done. I haven't put in the dialectical forms yet (although since that's pretty much dual oblique and dative plural, it won't take long), but I'd appreciate it if you could check over my work and make sure the Attic forms are all correct. ObsequiousNewt (εἴρηκα|πεποίηκα) 20:46, 6 May 2015 (UTC)

I have made some comments in your sandbox. — JohnC5 21:37, 6 May 2015 (UTC)

I found a new first declension type...
It appears that Νασικᾶς, ἀτταγᾶς, κατωφαγᾶς, ὀστρακᾶς, and maybe more are effectively 1st-M-alp-con-α-gen (-ᾶς, -ᾶ). Yeah... <tt>:V</tt> — JohnC5 07:32, 8 May 2015 (UTC)
 * This'll be Smyth 225, which I haven't implemented in any case. I'll admit, since it's only one differing form, there aren't very many such words, and I have a suspicion that all three of those are Doric words anyway, I neglected to implement it in favor of just using the GS parameter.
 * But I did a search, and found ἀρῠσᾶς (inscr. Delos), ἐμβᾰδᾶς (Theopomp. and Archipp.), φᾰγᾶς (Cratin.), φᾰκᾶς (Suid.), φᾰκῐνᾶς (BGU), φῐλοβορρᾶς (Hdn.Gr.), καταφᾰγᾶς/κατωφᾰγᾶς (various), κλειδᾶς (BGU), κορυζᾶς (Men.), λᾰχᾰνᾶς (Hdn.Gr.), μαμωνᾶς (Ev.Luc., an Aramaic loan), μασουχᾶς (Alex.Trall. and Paul.Aeg.) μαχαιρᾶς (POxy), ναμαρᾶς (Semitic), ὀδοντᾶς (Gloss.), ὀρνῑθᾶς (various papyri), παξαμᾶς (Gal. and Suid.), παστιλλᾶς (POxy and PIand), πελλᾶς (Hdn.Gr.), πῐνᾰκιδᾶς (Hdn.Gr.), πλᾰκουντᾶς (POxy), πορτᾶς (POxy), πραγμᾰτᾶς (Panticapæum), ῥᾰφῐδᾶς (PWürzb), σαγμᾰτᾶς (PFlor), σακκᾶς (Corcyrus and PLond), σαπουλανᾶς (Corcyrus), στροβῑλᾶς (POxy), σχοινᾶς (PSI and Gloss.), τᾰπῐτᾶς (POxy), τᾰρῑχᾶς (Stud.Pal.), ὑᾰλᾶς (IG), ζῡτᾶς (BGU). Interestingly, despite Smyth, the only word in LSJ with a genitive in -οῦ is κατωφᾰγᾶς (or -α-) and it lists both. Even Βορρᾶς, despite Smyth, lists only -ᾶ. A Perseus collections search yields neither Βορροῦ nor Βορρᾶ. This is about the point at which I'd go to consult another source, except that I'm not sure what that would be. ObsequiousNewt (εἴρηκα|πεποίηκα) 19:02, 8 May 2015 (UTC)
 * I looked through the proper noun section of Woodhouse and found these:
 * Ἀννιβᾶς, -ᾶ (Xen., Hell.), -οῦ (Plut.)
 * Ἀχιλλᾶς, -ᾶ
 * Βορρᾶς, -ᾶ
 * Νᾱσῑκᾶς, -ᾶ
 * Νομᾶς, -ᾶ
 * DGE has:
 * Ἀχιλλᾶς, -ᾶ
 * ἀρυσᾶς, -ᾶ
 * Βορρᾶς, -ᾶ (Lyr.Adesp.82, Cratin.222, PPetr.3.1.2.11 (III a.C.), UPZ 114.1.11 (II a.C.)), but -οῦ (Aristonym.7)
 * ἐμβαδᾶς, -ᾶ
 * It seems like -ᾶ is undoubtedly more common and Smyth was drunk when he wrote otherwise.
 * In related news, if you look at Ἑρμῆς, it and the -κλῆς nouns have a lot of dialectal endings.

It is accomplished
Or something. I don't think I missed anything—there are a couple inflections that Smyth shows like -οσ- or -ου- stems but they're like one or two words each. ObsequiousNewt (εἴρηκα|πεποίηκα) 18:46, 21 May 2015 (UTC)
 * Have you included the information from here? Also, have you added a notes section? Also, could we add comparative, superlative, and adverb parameters for the same of prettiness? Did we ever get the above -ᾶς, -ᾶ declension added? The other question is whether we might have the template create declensional categories like "AG dental stem nouns," "AG Doric declension nouns," etc. A Sorry to add more work but I wanted to wait till you were done before bringing these up.
 * It is also my intention, once this is ready and approved by other people, to begin replacing (and eventually deleting) all of the old tables. Any thoughts about this? — JohnC5 19:11, 21 May 2015 (UTC)
 * Also, congrats, looks great, I really appreciate it, you're a legend. — JohnC5 19:17, 21 May 2015 (UTC)
 * I think so; not yet; those are for adjectives which I haven't implemented yet either; I'll ask Textkit about that one; categories are easy enough; yes, I figured this template could replace all the old tables. I'll take care of the notes and do a bit of renaming tomorrow. ObsequiousNewt (εἴρηκα|πεποίηκα) 23:01, 21 May 2015 (UTC)
 * Yes...comparatives and superlatives are for adjectives...<tt>:V</tt> — JohnC5 00:48, 22 May 2015 (UTC)
 * Alright, I equalized the width. I'm going to be gone for something like the next three days, so I lay upon you the task (should you choose to accept it) of stress-testing that module. Find anything broken, anything missing, come back to me with a list when I return. ObsequiousNewt (εἴρηκα|πεποίηκα) 14:30, 27 May 2015 (UTC)
 * Howdy, I'm in the process of converting every AG noun I've ever added. The first part is in User:JohnC5/Sandbox4 and spills over into User:JohnC5/Sandbox1‎. There appear to be a few problems thus far.
 * I also feel bad about my words in that digamma-vau conversation. That was a bit meaner than I needed to be, but I was frustrated. I will, however, probably make a reference template for le Grand Bailly as it is a useful dictionary (though a later edition). — JohnC5 19:49, 29 May 2015 (UTC)

It is accomplished, again
And now begins error-checking time! Ignore ἀγνώς, though (that was just a test.) —ObsequiousNewt (εἴρηκα|πεποίηκα) 03:55, 3 July 2015 (UTC)


 * Oh wow, well done! I'm a bit busy nowadays, and need to cut down on my time editing Wiktionary; however, I'll do what I can, when I can to help with this. — I.S.M.E.T.A. 17:22, 3 July 2015 (UTC)


 * Well, don't feel obliged. John had been doing most of the testing before, anyway. —ObsequiousNewt (εἴρηκα|πεποίηκα) 18:06, 3 July 2015 (UTC)


 * Nice work. I'll be back in a few days and then I'll help out! — JohnC5 10:17, 4 July 2015 (UTC)


 * Okay, so, assuming that you haven't found any other obvious errors, I'll move the templates to the namespace.
 * Unless...
 * Okay, so there's an idea I've been kicking around, which boils down to this: instead of having the user try to remember eleventy template names, I write a bunch of regexes so that the user can enter in, as the only argument, the genitive singular (or, for adjectives, the (feminine and) neuter nominative singular). I don't think there are any pairs of declension types that share the two. Whaddaya think? —ObsequiousNewt (εἴρηκα|πεποίηκα) 16:32, 8 July 2015 (UTC)


 * you had me at eleventy. Do it. — JohnC5 16:37, 8 July 2015 (UTC)


 * Alright, it's done. There's one last thing which I was considering, which was to split the (1) first-declension stems in -ᾱ(ς)/-η(ς) into two separate tables (essentially as it was done in, although maybe not with a separate Epic table), and (2) third-declension stems in -ῐ(ς) into two separate tables—one in -εως for Attic, and one in -ῐος for pretty much everything else. This way the footnotes section isn't larger than the actual table. (Although with regard to (1) I don't remember if our policy is to include alternate inflections on the alternative form page or not.) —ObsequiousNewt (εἴρηκα|πεποίηκα) 21:24, 13 July 2015 (UTC)


 * It would seem correct to have the alternative lemma form on a different page... but please do split the -ῐ(ς) stems. I'm unsure what to do about the 1st-prx. — JohnC5 21:46, 13 July 2015 (UTC)


 * Alright, I've split the iota stems. I'll leave the ā-stems for now. Apart from that, though... I think we're ready for launch, once I write the docs. —ObsequiousNewt (εἴρηκα|πεποίηκα) 03:18, 14 July 2015 (UTC)

Sorry if that's all a bit piddling. I hope it has some worth. — I.S.M.E.T.A. 19:58, 21 July 2015 (UTC)
 * As I said, I don't really feel qualified to make comments on this, but I've got a few trifling ones to make anyway:
 * 1) Why are nouns' nominative and genitive singular forms, as well as adjectives' masculine, feminine, and neuter forms, separated by semi-cola, rather than commata? It's pretty unusual, and looks kinda strange; is there a reason for it?
 * 2) Adjectives' declension tables currently have thick vertical bars separating the singular forms from the dual forms and the dual forms from the plural forms. Can these be removed? Or, at least, the leftmost one? It looks strange that the case names' and the "Derived forms" row titles have a thicker right-side border than the "Dialectal forms" row title.
 * 3) Re  currently in User:ObsequiousNewt/Sandbox, the adverb is, not ; also, shouldn't the table list the various comparative and superlative forms, like the table at ἀγαθός currently does, rather than directing the reader to the usage notes?


 * I think (1) was a hangover from a few rapid changes we made... but it could also have been a suggestion from John to make things clearer. (2) I'll fix. (3) is, respectively, a typo, and the fact that ἀγαθός has a *lot* of comparative forms and I figured that would be the best way to solve it—but that's just a page for testing out the function in general; nothing there is necessarily how I think pages should look (I mean, I left off a lot of dialectical forms that I would otherwise have put there.) —ObsequiousNewt (εἴρηκα|πεποίηκα) 23:43, 21 July 2015 (UTC)


 * (1) I see. Well, if semi-cola are preferred to commata, how about we separate them with  (space-interpunct-space) instead? IMO, that looks better. What do you think, John? (2) Thank you; that looks a lot better. (3) OK, I understand. Thanks. — I.S.M.E.T.A. 18:11, 23 July 2015 (UTC)


 * Could you fix the module error (<font color="BB0000">Lua error in Module:grc-decl at line 31: no such inflection: μεθοδῐκοί ) currently in μεθοδικοί, please? — I.S.M.E.T.A. 22:11, 25 July 2015 (UTC)


 * Thanks for . I misunderstood how works; sorry. — I.S.M.E.T.A. 22:18, 25 July 2015 (UTC)