Google N-gram Corpus Counts
╭─ーー─┬───────────┬───────╮
│ 帰る │ 7,565,315 │ 98.0% │
│ 還る │   148,865 │  1.9% │
│ 歸る │     5,407 │  0.1% │ 🡠 sK (itaiji; not explicitly in refs)
│ 復る │       597 │  0.0% │ 🡠 adding (meikyo: 同じ道を引き返す意では「復」（「今来た道を復る」）が好まれる)
╰─ーー─┴───────────┴───────╯

@@ -15 +15,5 @@
-<ke_inf>&oK;</ke_inf>
+<ke_inf>&sK;</ke_inf>
+</k_ele>
+<k_ele>
+<keb>復る</keb>
+<ke_inf>&sK;</ke_inf>

I don't think 還る should be rK. I've seen that form numerous times. It's not rare in the way that 歸る is.

@@ -12 +11,0 @@
-<ke_inf>&rK;</ke_inf>

@@ -27 +26,0 @@
-<xref type="see" seq="1512150">返る・1</xref>
@@ -36,2 +35 @@
-<s_inf>of a guest, customer, etc.</s_inf>
-<gloss>to leave</gloss>
+<gloss>to leave (of a guest, customer, etc.)</gloss>

gg5, prog

I think this is clearer.

@@ -22,0 +23,2 @@
+<gloss>to create (an atmosphere, feeling, etc.)</gloss>
+<gloss>to produce</gloss>
@@ -24 +25,0 @@
-<gloss>to bring about</gloss>

Unidic/ngrams.

@@ -11,0 +11,3 @@
+</k_ele>
+<k_ele>
+<keb>醸しだす</keb>

G n-grams:
食らいつく	22987
喰らいつく	14148
食らい付く	 4069
喰らい付く	 3623
くらいつく	 9040

@@ -5,0 +6,3 @@
+</k_ele>
+<k_ele>
+<keb>喰らいつく</keb>

Unidic/ngrams.

Splitting. The kanji and readings are in two distinct groups.

@@ -5,4 +5,1 @@
-<keb>食い付く</keb>
-</k_ele>
-<k_ele>
-<keb>食いつく</keb>
+<keb>食らいつく</keb>
@@ -16,1 +13,1 @@
-<keb>食付く</keb>
+<keb>喰らい付く</keb>
@@ -19,6 +16,0 @@
-<reb>くいつく</reb>
-<re_restr>食い付く</re_restr>
-<re_restr>食いつく</re_restr>
-<re_restr>食付く</re_restr>
-</r_ele>
-<r_ele>
@@ -26,1 +17,0 @@
-<re_restr>食らい付く</re_restr>
@@ -33,0 +23,1 @@
+<xref type="see" seq="2797990">食いつく・くいつく・1</xref>
@@ -40,0 +31,1 @@
+<xref type="see" seq="2797990">食いつく・くいつく・2</xref>

Unidic/ngrams.

@@ -9,0 +9,3 @@
+</k_ele>
+<k_ele>
+<keb>食べもの</keb>

field tag should be sufficient

@@ -29,1 +29,1 @@
-<gloss>(astronomical) magnitude</gloss>
+<gloss>magnitude</gloss>

GG5, etc.

See rejected 2797930.

@@ -20,0 +20,5 @@
+<gloss>degree</gloss>
+<gloss>rank</gloss>
+<gloss>rating</gloss>
+<gloss>order</gloss>
+<gloss>points</gloss>

@@ -23,0 +23,1 @@
+<field>&astron;</field>

Yes, huge ngram count.

also, i think this should clearly be a (P) entry

@@ -8,0 +8,1 @@
+<ke_pri>spec1</ke_pri>
@@ -13,0 +14,1 @@
+<re_pri>spec1</re_pri>

gg5

splitting senses - change vs. revise

@@ -19,0 +19,3 @@
+<gloss>to transform</gloss>
+<gloss>to convert</gloss>
+<gloss>to turn</gloss>
@@ -20,1 +23,5 @@
-<gloss>to convert</gloss>
+</sense>
+<sense>
+<pos>&v1;</pos>
+<pos>&vt;</pos>
+<gloss>to reform</gloss>
@@ -23,1 +30,0 @@
-<gloss>to transform</gloss>

Bulk update of vi-vt tags from meikyo
-*- via bulkupd.py -*-

@@ -17,0 +18 @@
+<pos>&vt;</pos>

Agree. Not a good merge.

koj, daijr, daijs, gg5, prog, meikyo, shinmeikai

none of my sources merge these as the primary meanings are different.  does the adj-no belong on this one or the other one?  also, are n-grams truly indicative of adj-no use?  consider this non-adj 〜の example from eij:
 死体遺棄の疑い
allegations of abandonment of a corpse

@@ -8,3 +8,0 @@
-</k_ele>
-<k_ele>
-<keb>委棄</keb>
@@ -23,1 +20,0 @@
-<gloss>relinquishment</gloss>

ngrams

@@ -19,0 +19,1 @@
+<pos>&adj-no;</pos>

sK is not used for kanji forms that are in the major refs. Probably fine to hide 乗替え, though.

@@ -19 +19,5 @@
-<keb>乗りかえ</keb>
+<keb>乗り替え</keb>
+<ke_inf>&rK;</ke_inf>
+</k_ele>
+<k_ele>
+<keb>乗替え</keb>
@@ -23,5 +27 @@
-<keb>乗り替え</keb>
-<ke_inf>&sK;</ke_inf>
-</k_ele>
-<k_ele>
-<keb>乗替え</keb>
+<keb>乗りかえ</keb>

乗り換え	8889487	77.9%
乗換	2066936	18.1%
乗換え	403749	3.5%
乗りかえ	28550	0.3%
乗り替え	15050	0.1%
乗替え	1150	0.0%

Decluttering.

@@ -19,0 +20 @@
+<ke_inf>&sK;</ke_inf>
@@ -22,0 +24 @@
+<ke_inf>&sK;</ke_inf>
@@ -25,0 +28 @@
+<ke_inf>&sK;</ke_inf>

乗り換えする	7016
乗り換えし	15464

It's vs in shinmeikai.
I'll drop vs from the last two senses.

@@ -36,0 +37 @@
+<pos>&vi;</pos>
@@ -43,0 +45 @@
+<pos>&vi;</pos>
@@ -50 +51,0 @@
-<pos>&vs;</pos>
@@ -56 +56,0 @@
-<pos>&vs;</pos>

not vs in kokugos

Reverting.

@@ -14,0 +15,9 @@
+</r_ele>
+<r_ele>
+<reb>ほんとに</reb>
+<re_inf>&ik;</re_inf>
+</r_ele>
+<r_ele>
+<reb>ホントに</reb>
+<re_nokanji/>
+<re_inf>&sk;</re_inf>

Google N-gram Corpus Counts
╭─ーーー─┬───────┬───────╮
│ 本當に │ 5,073 │100.0% │
╰─ーーー─┴───────┴───────╯

see comments in the 1523060 entry
made ホントーに/ホントウに search-only
added 本當に

@@ -7,0 +8,4 @@
+<k_ele>
+<keb>本當に</keb>
+<ke_inf>&sK;</ke_inf>
+</k_ele>
@@ -13,8 +16,0 @@
-<reb>ほんとに</reb>
-<re_inf>&ik;</re_inf>
-</r_ele>
-<r_ele>
-<reb>ホントに</reb>
-<re_nokanji/>
-</r_ele>
-<r_ele>
@@ -22 +18 @@
-<re_nokanji/>
+<re_inf>&sk;</re_inf>
@@ -26 +22 @@
-<re_nokanji/>
+<re_inf>&sk;</re_inf>

getting kinda ugly...

ngrams

I noticed a NJS translator using "ホントに" in a post, so I checked the ngrams counts:
本当に	54156119
ほんとうに	1819224
ほんとに	6910762
ホントに	6252892
ホントーに	31974
ホントウに	19639

本とに gets 22k too, but I feel it's a step too far, unless we split off a "misspelling" version of the entry.

@@ -16,0 +16,12 @@
+<r_ele>
+<reb>ホントに</reb>
+<re_nokanji/>
+</r_ele>
+<r_ele>
+<reb>ホントーに</reb>
+<re_nokanji/>
+</r_ele>
+<r_ele>
+<reb>ホントウに</reb>
+<re_nokanji/>
+</r_ele>

really adj-f or just tending to form the first part of a compound noun?

GG5, etc. Ngrams

I'm changing my tune. I think the kana form really is more common.

@@ -23,0 +24,2 @@
+<pos>&adj-f;</pos>
+<misc>&uk;</misc>
@@ -25,0 +28,2 @@
+<gloss>rapport</gloss>
+<gloss>mutual touching</gloss>

Unidic/ngrams
https://ja.wikipedia.org/wiki/ふれあい

ふれあい in kana is by far the most common, but it seems to be because of it being a song title and an advertising slogan.

@@ -9,0 +9,3 @@
+</k_ele>
+<k_ele>
+<keb>ふれ合い</keb>

Ok. I'll do that.

@@ -21,7 +20,0 @@
-<reb>かなたこなた</reb>
-</r_ele>
-<r_ele>
-<reb>あなたこなた</reb>
-<re_inf>&ok;</re_inf>
-</r_ele>
-<r_ele>
@@ -45,4 +37,0 @@
-<stagr>あちこち</stagr>
-<stagr>あちらこちら</stagr>
-<stagr>あっちこっち</stagr>
-<stagr>アチコチ</stagr>

gg5, prog, daij, meikyo

I think かなたこなた/あなたこなた should be a separate entry. They're not vs and they only apply to sense 1. Splitting them out would allow us to drop the unsightly restr tag.

@@ -31,0 +32 @@
+<pos>&pn;</pos>
@@ -33 +34,2 @@
-<pos>&pn;</pos>
+<pos>&vs;</pos>
+<pos>&vi;</pos>
@@ -35,0 +38 @@
+<gloss>various places</gloss>
@@ -36,0 +40 @@
+<gloss>all over</gloss>
@@ -39 +42,0 @@
-<gloss>all over</gloss>
@@ -45,0 +49,2 @@
+<pos>&adj-na;</pos>
+<pos>&n;</pos>
@@ -46,0 +52 @@
+<pos>&vi;</pos>
@@ -48,2 +54,4 @@
-<gloss>to get things in the wrong order (back to front)</gloss>
-<gloss>to become muddled up</gloss>
+<gloss>muddled</gloss>
+<gloss>confused</gloss>
+<gloss>in the wrong order</gloss>
+<gloss>back to front</gloss>

彼方此方	30191
あちこち	3796605
あちらこちら	675093

@@ -6,2 +6 @@
-<ke_pri>ichi1</ke_pri>
-<ke_pri>spec1</ke_pri>
+<ke_inf>&rK;</ke_inf>

@@ -47,0 +48 @@
+<misc>&uk;</misc>

Meikyo vt and vi additions to n,vs entries
-*- via bulkupd.py -*-

@@ -15,0 +16,2 @@
+<pos>&vt;</pos>
+<pos>&vi;</pos>

GG5, etc. Unidic/ngrams

@@ -4,0 +4,3 @@
+<k_ele>
+<keb>食いだめ</keb>
+</k_ele>
@@ -13,1 +16,2 @@
-<gloss>stuff oneself with food</gloss>
+<gloss>stuffing oneself with food</gloss>
+<gloss>eating enough to go without food for some time</gloss>

okay

Some comments. The only resp. on the SLJ group was from Andreas Rusterholz, who is usually on the mark:
------------------------------
I think it's right to regard 変える and 替える as having different meanings. Nevertheless there are dictionaries (大辞泉 and others) which accept 見変える and 見替える as alternative reading of the same word. 広辞苑 and 岩波国語辞典 have only 見変える.

Regarding your question: The National Institute for Japanese Language and Linguistics has 移し替える as main entry and 移し変える / 移し換える as alternative writing (異表記). See http://vvlexicon.ninjal.ac.jp/db/ and search for うつしかえる。

http://csd.ninjal.ac.jp/comp/index.php is a Webデータに基づく複合動詞用例データベース with a lot of examples for 移し替える (2000)、移し変える (1618)、and 移し替える (494)。Looking at some of these examples I dont't think that it is justified to treat 移し変える and 移し替える as different entries. 
-----------------------------------------
I also raised it on TheJapanesePage, and a couple of NSJs responded:
---------------------------------
It's the same word in that it has the same basic root meaning and it is phonetically the same word (so you wouldn't necessarily think about which kanji so much if you heard it.) But if it was written down in kanji, the imagery would be appropriate to the situation. So I consider it the same word with two usages that require an entry in the dictionary.
--------------------------------------------
Let me try to explain the difference of those two in nuance.
But please understand that the interpretation might somewhat differ amongst age groups. Take my interpretation as one of “aged” people.

We use both of those kanji’s in our daily life like they are interchangeable, however, if you write:
この2つの箱の中身を移し変えてください。
I for one would see it a little bit weird. In my opinion, it should be 替 (or 換) because you want to 交換する of the contents of the two boxes.

It would be your point of view that counts. The kanji “変” has a strong association with “変化” whereas “替” (or 換) has with “交替” (or 交換). If you (only) see the (first) box, the “change” is to be done in its own “world”, however, if you see both of the two boxes, the “(ex)change” is to be done “between” the two. That’s the difference, in my humble opinion.

...Which means, well, my instinct tells me that 移し替える (or 移し換える) looks "more proper" in any case considering the meaning of the complex verb, and "変" would be "modern usage" and "acceptable", so to speak.
---------------------------------------
So what to do with it?
It's rather fine, but I suspect the way to handle it is to have two senses with "esp. ..." notes. I don't think they can be regarded as two distinct entries.

@@ -13,0 +13,3 @@
+<k_ele>
+<keb>移し換える</keb>
+</k_ele>
@@ -19,1 +22,3 @@
-<gloss>to shift (move) (an object) to (into)</gloss>
+<s_inf>esp. 移し変える</s_inf>
+<gloss>to shift something to</gloss>
+<gloss>to move something into</gloss>
@@ -21,0 +26,6 @@
+<gloss>to transplant</gloss>
+</sense>
+<sense>
+<pos>&v1;</pos>
+<pos>&vt;</pos>
+<s_inf>esp. 移し替える, 移し換える</s_inf>
@@ -22,1 +33,1 @@
-<gloss>to transplant</gloss>
+<gloss>to exchange</gloss>

Good point. JWN does distinguish between the two. I looked it up in the SketchEngine (which uses Unidic, and hence has them all together.) The concordance examples I went through look very similar, but then I may be missing subtleties.
I have no problem splitting these up with glosses as in Wordnet, but first I have asked about them on the SLJ group. The NSJs there may be able to help.

nikk

by having two different entries for 変える and 替える, we've implied that 変える and 替える are different.  does that mean that one of these should be marked iK?  nikk has only 移し替える

@@ -18,0 +18,1 @@
+<pos>&vt;</pos>

JWN,Unidic/ngrams

@@ -5,0 +5,3 @@
+<keb>移し変える</keb>
+</k_ele>
+<k_ele>
@@ -6,0 +9,3 @@
+</k_ele>
+<k_ele>
+<keb>移しかえる</keb>
@@ -13,0 +19,4 @@
+<gloss>to transfer</gloss>
+<gloss>to transpose</gloss>
+<gloss>to transplant</gloss>
+<gloss>to displace</gloss>

GG5

No, it's surely さかなりょうり.

@@ -13,0 +13,1 @@
+<gloss>fish dish</gloss>

nipp "魚料理（さかなりょうり）"

any source for it being read "ぎょ"?

@@ -8,1 +8,1 @@
-<reb>ぎょりょうり</reb>
+<reb>さかなりょうり</reb>
@@ -12,2 +12,1 @@
-<gloss>seafood dinner</gloss>
-<gloss>seafood-based cuisine</gloss>
+<gloss>fish-based cooking</gloss>

Entry created

Unidic/ngrams

@@ -6,0 +6,3 @@
+</k_ele>
+<k_ele>
+<keb>蒸しもの</keb>

Entry created

色むら	93687
色斑	2295
色ムラ	89343

@@ -7,0 +8,3 @@
+<keb>色ムラ</keb>
+</k_ele>
+<k_ele>
@@ -11,0 +15,6 @@
+<re_restr>色むら</re_restr>
+<re_restr>色斑</re_restr>
+</r_ele>
+<r_ele>
+<reb>いろムラ</reb>
+<re_restr>色ムラ</re_restr>

GG5, Unidic/ngrams

@@ -4,0 +4,3 @@
+<k_ele>
+<keb>色むら</keb>
+</k_ele>

Entry created

larvae = pl.

@@ -23,1 +23,1 @@
-<gloss>any moth whose larvae feeds on fruit</gloss>
+<gloss>any moth whose larvae feed on fruit</gloss>

GG5, ルミナス (codling moth larva), Daijr (more general)

@@ -19,0 +19,4 @@
+<gloss>codling moth larva</gloss>
+</sense>
+<sense>
+<pos>&n;</pos>

(meant to say nipp)

nikk

mk

@@ -12,0 +13 @@
+<pos>&vt;</pos>
@@ -20,0 +22 @@
+<pos>&vt;</pos>

Agree. Not a good merge.

gg5, prog, koj, daij, meikyo, shinmeikai

splitting from 1587090

ＫＯＤ追加語彙 has:
等級別料率制度 【保険】 〔自動車保険の〕 an insurance premium rating system.

I think it's more the rating rather than the points which lead to it.
I'll amend the 等級 entry.

Google for "自動車保険 等級" and lots of insurance related sites will appear discussing this

GG5, etc.

新和英中辞典 nikk prog

I think so. Very common.

研究社 英和コンピューター用語辞典 example
eij example

maybe worth having? we already have 予告なしに

OK

See my 2nd edit.

@@ -6,0 +7,3 @@
+<k_ele>
+<keb>駅チカ</keb>
+</k_ele>
@@ -8,0 +12,5 @@
+<re_restr>駅近</re_restr>
+</r_ele>
+<r_ele>
+<reb>えきチカ</reb>
+<re_restr>駅チカ</re_restr>

I'll go back to the original and if we want to change it we can start again.

@@ -7,3 +6,0 @@
-<k_ele>
-<keb>駅チカ</keb>
-</k_ele>
@@ -12,9 +8,0 @@
-<re_restr>駅近</re_restr>
-</r_ele>
-<r_ele>
-<reb>えきチカ</reb>
-<re_restr>駅チカ</re_restr>
-</r_ele>
-<r_ele>
-<reb>エキチカ</reb>
-<re_nokanji/>
@@ -27,5 +14,0 @@
-</sense>
-<sense>
-<pos>&n;</pos>
-<s_inf>usu. 駅チカ</s_inf>
-<gloss>shopping arcade near a station</gloss>

It's probably better to keep them separate actually as 2nd sense can probably be written "駅地下" while it's not written as "駅近"

エキチカ	1856

OK. Maybe this?
I'll flag 2797970 for deletion.

@@ -17,0 +18,4 @@
+<r_ele>
+<reb>エキチカ</reb>
+<re_nokanji/>
+</r_ele>
@@ -22,0 +27,5 @@
+</sense>
+<sense>
+<pos>&n;</pos>
+<s_inf>usu. 駅チカ</s_inf>
+<gloss>shopping arcade near a station</gloss>

Cancelled

Approve if the amendment to 駅近 is approved.

ＫＯＤ追加語彙

ngrams (88k)

I guess that works.

Twitter usage

駅地下	36015
駅チカ	88041

@@ -6,0 +7,3 @@
+<k_ele>
+<keb>駅地下</keb>
+</k_ele>
@@ -8,0 +12,5 @@
+<re_restr>駅チカ</re_restr>
+</r_ele>
+<r_ele>
+<reb>えきちか</reb>
+<re_restr>駅地下</re_restr>

to make it clearer this is not from 近 (but 地
下)

@@ -12 +12 @@
-<gloss>underground shopping arcade near a station</gloss>
+<gloss>underground shopping arcade connected to a station</gloss>

ＫＯＤ追加語彙

ngrams (88k)

"to be made to resign" sounds like there are people with cudgels waiting there.

@@ -17 +17 @@
-<gloss g_type="expl">rooms where employees who cannot be fired are sent to be made to resign</gloss>
+<gloss g_type="expl">rooms where employees who cannot be fired are sent to encourage resignation</gloss>

https://www.nytimes.com/2013/08/17/business/global/layoffs-illegal-japan-workers-are-sent-to-the-boredom-room.html

@@ -16 +16,2 @@
-<gloss g_type="expl">rooms where employees who cannot be fired are sent</gloss>
+<gloss>boredom room</gloss>
+<gloss g_type="expl">rooms where employees who cannot be fired are sent to be made to resign</gloss>

Daijr

Please put long explanations as a comment

@@ -12,1 +12,5 @@
-<gloss>Expulsion room. Banishment Room. In companies it refers to rooms where employees who can not be fired are sent to, with nothing to do, humiliated, until they decide to leave the company. Also, they are more often given never-achievable, heavy work quota on the contrary, and demanded by the managers to achieve it, or resign the company.</gloss>
+<xref type="see" seq="2643660">追い出し・おいだし・1</xref>
+<misc>&col;</misc>
+<gloss>expulsion room</gloss>
+<gloss>banishment room</gloss>
+<gloss g_type="expl">rooms where employees who cannot be fired are sent</gloss>

chuujiten

@@ -37,2 +37,3 @@
-<gloss>to get one's teeth into (metaphorically)</gloss>
-<gloss>to get to grips with</gloss>
+<gloss>to jump at (e.g. an offer)</gloss>
+<gloss>to snap at</gloss>
+<gloss>to get one's teeth into</gloss>

G n-grams:
食いつく	79487
食い付く	11321
食付く	  314
喰いつく	 5488
喰い付く	 2908
喰付く	   33

The 喰～ forms are not in my refs (gg5, koj, daijs), but they are quite common

@@ -11,0 +12,9 @@
+</k_ele>
+<k_ele>
+<keb>喰いつく</keb>
+</k_ele>
+<k_ele>
+<keb>喰い付く</keb>
+</k_ele>
+<k_ele>
+<keb>喰付く</keb>

Interesting glossary.

http://www.toraya-group.co.jp/siteinformation/wor_index.html

Flour used for making higashi, small traditional sweets 
severed with Japanese matcha tea. As the flour has already 
been cooked, it is simply mixed with fine damp sugar then 
pressed in moulds before serving.

Daijr, http://www.toraya-group.co.jp/siteinformation/wor_index.html