All of my previously cited sources
Also https://ja.wikipedia.org/wiki/刀

>I will reject this because no sources have been referenced. I'm sure the entry can be improved, but I don't feel Alan's edit is a good 
springboard for that.

The sources were all the ones I had previously cited!
Why should I have to re-cite all of them, every single time, when I have already cited them?

>世界大百科事典 第２版 katana:

katana? Not かたな?
That would suggest it is talking about the English word, which strictly refers to 打刀.
As it is not the same word, that source is clearly completely irrelevant.

>daijr says 細長い (2ア) and 長い (2イ).

(2ア) and (2イ) indicates that these are sub-senses.
Not what the term exclusively refers to.
Note that daijr also states "3) 小さい刃物。".
Did you choose to ignore that, because it conflicts with (utterly destroys) your position?

…and what about how koj says "1) 刀身が短い片刃の刃物。"
Not to mention "3) 小さい刃物。きれもの。" (this is also identical to sense 4, in daijs)

As for it referring to Japanese swords, specifically…
Well, I could find a whole bunch of different sources, but would Japanese Wikipedias article, on かたな suffice?
The first image you see, straight away, is a French military sabre …and it goes on to list tons of non-Japanese blades.

>/…/短刀など/…/

短刀 is typically a knife.
Note your other source (nipp katana), which states "30センチメートル未満のものを短刀とよび分ける".
Not that there is any kind of lack of sources, that refer to 短刀 as a knife.

@@ -21,2 +21 @@
-<gloss>Japanese sword (esp. a single-edged one)</gloss>
-<gloss>katana</gloss>
+<gloss>(single-edged) blade (usu. sword or knife/dagger)</gloss>

(nipp sekaidaihyakka)

@@ -21 +21 @@
-<gloss>sword (esp. a single-edged one)</gloss>
+<gloss>Japanese sword (esp. a single-edged one)</gloss>

brit: katana entry:
太刀 (たち) ，打刀 (うちがたな) ，脇差 (わきざし) ，短刀などの日本刀の総称。

世界大百科事典 第２版 katana:
広義には日本刀全体を指すこともあるが，狭義には刃を上向きにして腰にさす2尺(60.6cm)以上の刀身のもののみを刀という。

nipp katana: 
ものを切るのに適したように片刃にこしらえた武器。刺突用に両刃にこしらえた武器は剣という。刀身の長さが60センチメートル以上のものを太刀(たち)・刀、30
～60センチメートル未満のものを脇差(わきざし)、30センチメートル未満のものを短刀とよび分ける。

koj, daijs, daijr

I suggest removing the dagger/knife sense - it seems to be arch and just risks making the entry more confusing than it has to be.
Saying sense 1 often refers to "knives/daggers" (as Alan did in a rejected suggestion) doesn't match with what's in the references - nipp 
and sekaidaihyakka both say 60cm+. daijr says 細長い (2ア) and 長い (2イ). 
All senses that suggest it's something shorter seem to be archaic.

@@ -21 +21 @@
-<gloss>(single-edged) sword</gloss>
+<gloss>sword (esp. a single-edged one)</gloss>
@@ -23,5 +22,0 @@
-</sense>
-<sense>
-<pos>&n;</pos>
-<gloss>dagger</gloss>
-<gloss>knife</gloss>

sense 5 - not a Japanese concept, doesn't need an expl.

@@ -46,2 +46 @@
-<gloss>knife money</gloss>
-<gloss g_type="expl">knife-shaped commodity money used in Zhou dynasty China</gloss>
+<gloss>knife money (knife-shaped commodity money used in ancient China)</gloss>

I will reject this because no sources have been referenced. I'm sure the entry can be improved, but I don't feel Alan's edit is a good 
springboard for that.

>I don't think the romanized "katana" should be in a different sense; they're the same thing.

"(single-edged) sword" includes falchions, sabres, cutlasses, hangers, Messers, dussack, backswords, kopis, all kinds of Dao, shamshir, saif, tulwar, Kilij, shotel etc etc …and also katana, tachi (大刀), tachi (太刀), nodachi/oodachi, nagamaki, kodachi, wakizashi…

So…
Clearly not the same sense.

Not unless you propose making it something like
"(single-edged) sword; katana; wakizashi; tachi; nagamaki; dao; sabre; Messer; falchion; dussack; hanger; backsword; scimitar"
Note that the above list is by no means exhaustive, so there are types missing. Also I used the umbrella term scimitar for all the distinctly different types of Middle Eastern/Indian blades …and using tachi to cover oodachi/nodachi/kodachi as well (as well as 大刀), and the general term Dao, to cover all of the countless different types of single edged Chinese swords
…and yet, it still looks rather preposterously long.

@@ -21,7 +21 @@
-<gloss>(single-edged) sword</gloss>
-<gloss>katana</gloss>
-</sense>
-<sense>
-<pos>&n;</pos>
-<gloss>dagger</gloss>
-<gloss>knife</gloss>
+<gloss>(single-edged) blade (usu. knife/dagger or sword)</gloss>

sense 5 - not a Japanese concept, doesn't need an expl.

@@ -46,2 +46 @@
-<gloss>knife money</gloss>
-<gloss g_type="expl">knife-shaped commodity money used in Zhou dynasty China</gloss>
+<gloss>knife money (knife-shaped commodity money used in ancient China)</gloss>

GG5, etc. Koj, Daijr/s

I don't think the romanized "katana" should be in a different sense; they're the same thing.
I wouldn't mind if "dagger; knife" went into the first sense. They were there originally, but in 2007 Rene split them on the grounds that their usage was "arch" (we might say "hist" now.) Then I dropped the "arch" in 2018.

@@ -22,3 +21,0 @@
-</sense>
-<sense>
-<pos>&n;</pos>

As I have come to understand, JMdict indicates that descriptions separated by a semicolon, within the same sense, indicates that the separated descriptions are different ways of describing the same thing.

Also, shouldn't the senses of "(single-edged) sword" and "dagger; knife" simply be in a single sense of "(single-edged) blade"?
Or "(single-edged) blade; (single-edged) sword; knife; dagger"?

…and when speaking of かたな, in terms of a Japanese sword, is is sometimes used to refer to "Japanese sword", generally, and sometimes as katana (in the English sense of the word. What is known as 打刀, in Japanese), specifically.

Note that neither Japanese, nor Chinese, have the same knife/dagger vs sword distinction, as exists in English.
They merely have terms for blades, distinguishing by number of edges, but not length. (though that distinction has become somewhat blurred, in Japan …after they stopped making double-edged swords or combat knives/daggers, and tool knives are all single edged, aside from a few very specialized exceptions. Hence meaning that the distinction had little practical value anymore, as practically all blades were single edged. Until Modern times, that is)
This leads to a lot of erroneous translations. (arguably _especially_ by Japanese people)

@@ -21,0 +22,3 @@
+</sense>
+<sense>
+<pos>&n;</pos>