Are these two rapier swords one and the same?

Pikka

There are two very similar entries:
Minifigure Rapier (93550) and Sabre (37846) - image attached.
The one that I own has "37846" engraved under the hand guard.
Should they be merged into a single entry with two Design IDs?

mfav

Please, no. If it ships from LEGO with a part number stamped on it, let it stand in the catalog with that part number. The catalog is messy to begin with because LEGO makes a part with one ID, then remolds the part differently and ships it with the same ID. They'll also make a part with one ID, remold it the same, and ship it with a different ID.

I have a related rant here: https://hispabrickmagazine.com/blog-en/blog-en.php?id=33502

If we "merge" 93550 with 37846 now…this being an arbitrary decision for reasons which may be well-intentioned…ten years from now when somebody wants a 93550 and gets shipped a 37846, then the seller potentially has a problem that didn't need to be a problem.

I happen to run into this problem fairly regularly. It's a pain for me to have to explain these minor differences over and over to sellers when I want one specific part and get a similar part…all because the catalog sees them as one thing and not two.

Take a look at wizard hats
http://v4ei.com/brickref/wizard-hats.php

and bars with clips
http://v4ei.com/brickref/48729v41005.php

I realize not everybody is the part geek that I am, but for some of us these differences can be important in both terms of collecting and building.

Hoddie

The key question for me is always "are they functionally different," though I know many people who believe as @mfav does. When putting together a vintage set, getting the correct part variant is incredibly important for some.

In the case of these items, they should at the very least be linked by a mold variant relationship so that buyers can easily see each exists separately.

Pikka

@mfav Our store doesn't differentiate between mold variation, not because we don't want, but it's hard to keep record of all those minor differences. Nevertheless, if a buyer asks for a specific mold, it's not a problem to check if we have the right one.
I examined those two swords again and indeed there are some minor differences – not functional ones – so they might be linked by a mold variant relationship, as @Hoddie suggested. I checked BL, Rebrickable and Peeron they have only one entry (93550).

mfav

@Pikka I know the arguments. I know the sellers need a new box for every version. That takes space. And so on. Each AFOL has their own way of doing things.

I'm arguing for future-proofing and the unknown. It may be that in seven years the green 93550s spontaneously burst into flame when they encounter sunlight after midnight…or who knows what. Next year there may be another rapier with a new number made of a new plastic derived from the tears of old resellers combined with dashed expectations and ambitions, and this plastic may be more rubbery or more stiffy or I don't know what. That version may become much valued or much despised.

There will continue to be changes in plastic formulation, color formulation, new colors, and so on. My argument is really more for a robust and flexible database that contains correct information as opposed to a janky jumbled-up database where umpteen different versions are convoluted and conflated with one another. The data should be as pristine as possible.

Functionally different, functionally indifferent…these are not inherent attributes, but relative ones. The function or lack of function requires a relationship to another part or to a process. I seek objective classification of the part within the database, not subjective classification.

I don't know how similar parts should be related to one another specifically. But I feel that if there are two different part numbers, then there should be two different records. Within the individual records there can be a field that relates one to the other. This is as opposed to one record containing two (or more) parts.

If I search for 93550 and end up on a page with 37846, that looks like an error. I might search for 93550 and not find it because in the catalog it is 37846 with a note about being 93550 also. These scenarios are subject to the database structure and the programming that coincides, but I don't see any benefit in "merging" different numbers.