Are underside reinforcements relevant?

[Peter Blomberg]

Dear LDraw community,

Part moulds change over time as LEGO improves their designs. This leads to parts having relatively minor differences in their shapes. Sometimes this affects functionality, but often not.

When it comes to underside reinforcements, is it a detail that can be ignored or would you like to see all individual variants for the same part modeled in LDraw? Is it sufficient for you to have one variant in your digital part library or would you like to have all variants?

[Chris Böhnke]

Really hard to say...

Probably this depends on the part itself? Though making such differences makes it a bit pointless.

There are a few builds which have the underside exposed, so having the 'correct' era brick available might be beneficial here.

On another note, some parts (especially the most old and basic ones) have such a confusing array of underside differences, most sites seem to struggle keeping up with just catalogueing them. Also how do you determine which is the correct version for a specific printed brick?

Another issue is probably finding parts. Imagine you have like 3001a.dat, 3001b.dat, etc... Picking the right brick (or at least the same brick everytime) can become quite annoying in some programs.

In terms of 'realism' I would say 'yes', but in terms of practically I tend towards 'no'.

[Gerald Lasser]

Interesting question.

concerning mould differences, e.g. there are different moulds of 3001, where the walls got thinner and to still have cluth-power, some very small filltes have been added. those are not modeled, we kept the 4 LDU thick part. That is fine

Whenever an understud is held in place (reinforced) with a fillet, I would tend to have this modeled. As it is also part of the geometry that prohibits a bar or something else going into this place.
If such a thing is completely invisible, as it is part of an animal consisting of two distinct halfs, I would leave it up to the author, but those can be skipped IMHO, if it is not a transparent part, e.g. like the deer.

An example for a considerable mould change here is the "Duplo Roof Sloped 33  4 x  4 with  9 Bottom Tubes with Awning Overhang". This has two moulds (and design-IDs), which are for the basic Duplo build practically the same. However their underside is very different. 9 versus 5 anti-studs. So this is definately a "yes" to be modeled"

https://library.ldraw.org/tracker/24487
https://library.ldraw.org/tracker/24647

Examples where the reinfocement could be skipped are the "blates" the two plate thick "baseplates" those have usually some diagonal reinforecements 20-30 LDU wide and 1 LDU high, so basically they just make the top of the plate thicker, those are not necessary in my view. Mecabricks models them, but those are not necessary IMHO.

BR
Gerald

[N. W. Perry]

I envision underside details like this to be part of a system of aesthetic enhancements that can be applied client-side, just as primitive substitution is currently. I would imagine it being settable on a per-part basis, so that you could apply underside detail only to those elements that are exposed in your build, for example, or if you want to choose a certain historical variant of a printed brick.

So you would not have to model endless versions of library parts, but rather code the information as some kind of metadata or perhaps a shadow library. Only the base version would need to appear in the library.

This is one of those things that I envision as part of LDraw 2.0 (which, to be clear, has yet to exist).

[Peter Blomberg]

I envision underside details like this to be part of a system of aesthetic enhancements that can be applied client-side, just as primitive substitution is currently. I would imagine it being settable on a per-part basis, so that you could apply underside detail only to those elements that are exposed in your build, for example, or if you want to choose a certain historical variant of a printed brick.

So you would not have to model endless versions of library parts, but rather code the information as some kind of metadata or perhaps a shadow library. Only the base version would need to appear in the library.

This is one of those things that I envision as part of LDraw 2.0 (which, to be clear, has yet to exist).

What about separating the underside variations into subfiles, leaving only the parts common to all variations in the main file? Complete variations could then be constructed using the shortcut approach.

The part-independent common reinforcements could also be made into primitives in the same way technic has many modular primitives to choose from. That would make the creation of mold variations a matter of mix and match.

[Willy Tschager]

With so many parts still missing in the library I prefer new parts over variants.

w.

[N. W. Perry]

What about separating the underside variations into subfiles, leaving only the parts common to all variations in the main file? Complete variations could then be constructed using the shortcut approach.

The part-independent common reinforcements could also be made into primitives in the same way technic has many modular primitives to choose from. That would make the creation of mold variations a matter of mix and match.

That's already essentially the practice. Someone still has to author those subfiles/prims, and then all the different part variants, and they have to go through the review process.

The difference in my imagined method is that you only ever have to author one basic version of the part. Then any variants like underside supports, solid/hollow/blocked/vented stud, etc., could be "turned on" in your editing program, or embedding them in a meta- or shadow file, like how LDCad does with snap info.

With so many parts still missing in the library I prefer new parts over variants.

I would agree, given the need to choose. However, new parts often tend to be fairly intricate and may be out of reach for less experienced part authors. Working on something like underside variants could allow something for those authors to work on who aren't feeling up to the big stuff.

[Orion Pobursky]

As with all things, the library is a tradeoff between specificity (true to the real) and practicality (ease of creation, visibility in a model, etc). 

We tend to draw the line at "functional" elements of a part. We omit model lines and injection points. Other things line underside splines and dimples that wouldn't really be visible in a model. We do have variants for part whose outward appearance is different or if a hole/connection is changed such that other parts fit/don't fit. 

Since what to include is somewhat subjective, there aren't any hard and fast rules. Typically you should ask yourself:

- Does this variant provide any significant visual change during normal part usage?
- Does this variant provide or remove a connection point for other parts?
- Is this variant recognized as distinct (i.e. not group with an existing part) by the wider parts cataloging community?

If the answer is yes to any of these, then a variant *might* be warranted.

Sometimes the decision is clear, sometimes it's not. This is why we have forums for discussion and a Parts Library Administrator to make decisions. As parts authors and reviewers, we should be having healthy debate and be accepting of the consensus/admin even if the decision isn't in our favor.