User talk:Qorilla/Conjugation templates

Discussion
I am moving the discussion here. I read the guide and am trying to understand the system. Here is what I understand so far:
 * 1) There are eight templates for the front vowel verbs:
 * hu-conj-ek-t
 * hu-conj-ek-ett-el
 * hu-conj-ek-ett-sz
 * hu-conj-ek-ett-esz
 * hu-conj-ek-szem
 * hu-conj-gyek
 * hu-conj-ek-stem-tem
 * hu-conj-ek-stem-ettem
 * 1) This will be repeated for the back vowel words, so a total of 16 templates.
 * 2) All current hu-conj-xx templates will be replaced by the new ones. Who will do the work?
 * 3) The current hu-conj-ek will be replaced by two new templates: hu-conj-ek-t and hu-conj-ek-ett-esz. Why two when we can have one? We only need to add the past tense as a parameter just like in the current template.
 * 4) Recommendations for name change to make it easier to figure out the template from the template name:
 * hu-conj-ek-sz-t (kérek, kérsz, kért)
 * hu-conj-ek-el-ett (nézek, nézel, nézett)
 * hu-conj-ek-sz-ett (tépek, tépsz, tépett)
 * hu-conj-ek-esz-ett (keltek, keltesz, keltett)
 * hu-conj-ek-szem (no change)
 * hu-conj-gyek (no change)
 * hu-conj-ek-el-ett-shortstem
 * hu-conj-ek-esz-ett-shortstem
 * 1) There is a problem in the vedlik table,vedlelek should be vedellek.
 * 2) Instead of the text-like written guidelines, a table would be more helpful. List each template in the first column, present a few examples in the second column, list the optional parameters in the third column. Possible comments in the 4th column.

These are my observations so far. I hope you won't take my comments as criticism. They are not. It's great that you put time and effort into this challenging task. Have you had a chance to discuss it with others? --Panda10 22:12, 20 July 2010 (UTC)


 * 1) Yes
 * 2) Also for the rounded front, so 24 templates.
 * 3) Well I think we should leave the currently used templates as they are, so existing articles are not disturbed and gradually change them to the new version, and when adding new verbs we would use new templates.
 * 4) hu-conj-ek-t and hu-conj-ek-ett-esz are very different:
 * 5) * ek-t has no bouning vowel, while ett-esz has always (kérni, kérnél, rejteni, rejtenél.
 * 6) * The ette option can only appear in -ett-xx verbs, never in -t.
 * 7) * t -> ts and szt->ssz (e.g. bánt bántson - akaszt akasszon) only happen in -ek-ett-esz
 * 8) * We could however unite ek-t and ek-ett-sz (kér and lép), and signal the difference with a flag, but it would break the logic of system and the two have to be differentiated anyway, if not in the name then in a flag.
 * 9) Well, it's not logical for me. The first thing to decide about a word (after backness) is how the past is formed. We only need to proceed to the second person if it is -ett. We should go from outer category to inner. Also, then we wouldn't need to say -el-ett, because -el implies -ett, just like -esz. But we can rename the order if people think it's better so. But it still won't be totally intuitive, as the last is past, while the first two are present.
 * I found that the short stem categories depend on another form: néztem (see details below). It is used as an easily testable 'indicator' (there are many such background implications and the names don't tell every difference, as some implied features are automatically in the template if it is implied by the combination of template name, options, flags, last letter of the verb, etc). For example if it uses -ettem it means the verb uses the short stem whenever possible, if it uses -tem, it shortens less often.
 * 1) That might be true, but the point is that I made it so that some verbs simply don't have the long form, and have a blank cell whenever it is not possible to use the short stem (like all subjunctive). I was thinking about the -lak form and thought my solution is a possibility to make use of the short stem (if we accept that vedlik has no long stem as a premise, than we don't have much better choices). But in reality the form vedllek/vedellek/vedlelek is nonexistent anyway. It would mean you are a dog and you talk to your hair that you will get rid of it... Quite an exotic situation. On the other hand we could of course also squeez out the long form from all verbs like sikoljon, rejeljen (rejlik), feseljen (feslik), vedeljen, csukolj, but I wanted to avoid that after talking to people, who all felt there is no long form for these verbs. This is a rather minor question as I there are very few such transitive verbs where the -lak form is possible at all. One other is sínylik, megsínyli, but saying *megsínyellek is not likely, as you would be talking to an inanimate concept.
 * 2) The text guidelines were just temporarily drafted for your question, I plan to create a table, and maybe a diagram/chart.

I haven't discussed it yet here in Wiktionary. I made it so 95% ready like a year ago. That was quite a tedious work. First I did it in a fewer number of templates, but people here in WT said that it would be too much effort for the server to deal with all the if-else decisions. So then I split them to this version, experimenting with more variants, drawing tables and diagrams of how the features are related and which implies which, etc. Then I just left it and archived it because it took so long and was hard to handle like 25 templates in parallel and I was still not finished. But now summertime again, so I just thought why let it rot on my hard disk if only so little needs to be done. So that's why I haven't discussed it recently: because it just wasn't finished. Qorilla 23:49, 20 July 2010 (UTC)

I created a chart. I'll make a table as you suggested, a bit later. Qorilla 13:20, 21 July 2010 (UTC)

The stem shortening verbs: We could remake it to correspond to be named as the bold form + "shortstem". But we will need +2 template for each backness so a total of +6 templates would come into play. That's why I think it's more practical as it is now.


 * -ettem (rejlik) (-ett-esz)
 * -tem
 * past with -t (perel, szerepel) (-t-sz)
 * past with -ett (default) (képez, inog [can't find equivalent in front vowel. any verb is good that doesn't end in z,sz,dz, etc.] ) (-ett-el, -ett-sz) Qorilla 17:52, 21 July 2010 (UTC)
 * It's ok to leave everything as you planned. I have not looked into verb structures as deeply to say anything. Go ahead and implement it, along with the appropriate Documentation pages. Thanks. --Panda10 20:34, 21 July 2010 (UTC)
 * Sorry if I seemed like not accepting criticism, it's not easy for me to write English and I may sound harsh sometimes. I don't say I fully understand all these structures at all. You are a very productive contributor to Wiktionary so your opinion does matter. Qorilla 21:08, 21 July 2010 (UTC)
 * I think your English is excellent. And I was not angry at all when I replied. I really mean that you should go ahead and implement it. I simply don't have a better suggestion at this time. I still don't understand the template selection fully, but I'm sure eventually I will. --Panda10 21:23, 21 July 2010 (UTC)