Actually, I never said the mxs didn't match the catalog number. I said they weren't from the same session.The notes as they were originally written implied that they were by giving only one date and not specifying that it applied only to Side A. In that regard I was correct, as fixbutte's additional information and amended notes make clear. Your contributor would have been correct if his disc had been as described on the ODP, which leads me to suspect that he might have consulted the ODP and failed to notice that one of the sides listed there did not match the ones on his record.
I cannot say that you are completely wrong with regard to the ODP's flaws, but I would like to say a few words in its defense. When Steve Abrams set out to create the listings that are the foundation of the ODP, he attempted a gargantuan task at a time when the available information and tools were nothing like what they are today. To get where he wanted to be in something less than a lifetime, he used sources that were inherently riddled with misinformation and ambiguity, notably auction lists (Picture the old Record Research auction lists, if you know them, and you will know what that means.) Moreover, to make the data fit into the earliest electronic vehicles, it had to be somewhat adulterated with abbreviations and other conventions. The worst flaws in the ODP therefore pop up in the form of inexact artist credits and titles. Abrams was a competent discographer and usually paired his slap-dash list of issue numbers and somewhat correct artists and titles with spot-on information about mxs, and recording dates. What he created was, and still is, remarkable for its scope and degree of completion. Not even this site, for all its energy and modern advantages, does what the ODP does, or can really hope to ever do what it does, without years of further effort and the contributions of many, many people. I should also point out that without the constant attention of numerous moderators, this site would be as riddled with inaccuracies as the ODP, if not more so.
Ty Settlemier, who took it upon himself to bring the ODP to the online world, is well aware of the issues the ODP has. He inherited most of them from Abrams. Ty is an exacting scholar, and probably more ambitious than is good for him. He has no editorial team or assistants, but he continues to expand the ODP with the best available information while trying to identify and correct the errors that are there. He has stuck with it for longer than anybody else I know of online. He is readily accessible and welcomes help in improving the accuracy of his site. The ODP is better now than it was when I first made Ty's acquaintance nine years ago. It will continue to improve while Ty runs it. I encourage anybody who finds an error or a discrepancy in the ODP to bring it to Ty's attention. If enough of us are part of the solution, eventually there will no longer be a problem.
It is also important to point out that the 78 discography site should never be used as a strong/ rock-solid source of information as it quite often is just plain wrong or misleading in the info that it presents. One should should always compare the data presented there with other sites when attempting to make a decision about things such as issue date to ensure its reliability.
All your reasoning about matrix numbers not matching the catalog number would be appropriate if it was an original release. Decca's 25,000 series, however, was reserved for album reissues when it started at the end of 1946. And indeed, "The Dipsy Doodle" (mx. 62760) b/w "That's a-Plenty" (mx. 62761) was first issued in December 1937 on Decca 1553. "Canadian Capers" (mx. 69979), in contrast, was first issued as the B-side of "The Whistler's Mother-In-Law" (mx. 69978) in January 1942 on Decca 4118 (both sides recorded on December 1, 1941).
In addition, the Online Discographical Project (78discography.com), which has the original Decca records correctly, is completely fragmentary and often wrong with the first numbers of the 25,000 series. Actually the Milt Herth Trio album A-539 (with individual records 25064, 25065, 25066, and 25067) had the following sides, as found on a Russian site:
A Rockin' In Rhythm
B Honky Tonk Train Blues
C The Dipsy Blues
D Canadian Capers
E In An 18th Century Drawing Room
F Toy Trumpet
G The Whistler And His Dog
H Down South
Sides C and D (or 3 and 4) form Decca 25065 as listed here. "That's A Plenty" obviously was never part of the album.
I do find that somewhat strange. Not unheard of, but certainly unusual. To deliberately create two discs with the same number and similar but still different contents at the same time is bound to have its drawbacks. I haven't seen the version of 25065 listed in the 78 discography, but assuming that it does exist and that's why somebody listed it, I would guess, as you did, privatecitizen, that it pre-dated the album and that the version with "That's A Plenty" was withdrawn to make room for the album version. It was probably a choice between that or having two different numbers extant at the same time with a shared A-side.
Side B is separated from side A by over 7000 mx numbers. It certainly was not recorded at the same session or on the same day as side A. Rust did not consider Canadian Capers worthy of inclusion in Jazz Records, or else it fell after the 1942 cut off date, which is also possible. The last Herth record he lists had a 67000 matrix and was recorded in 1940.