Minor changes to the Official Library Spec

[Orion Pobursky]

Based on feedback, I'd like to proposed a few minor changes to the Official Library Spec:

Clarification to Sticker Geometry:
- For stickers printed on a transparent medium, the type 1,3, or 4 lines that comprise the pattern shall be preceded by a 0 BFC NOCLIP statement and ended by a 0 BFC CLIP statement
- The pattern should be oriented such that the top of the pattern aligns with the +Z axis. Exceptions to this rule may be made on a case by case based.

Claification to Sticker Shortcuts:
- Require a !KEYWORDS entry with text similar to Pattern !KEYWORDS Entry:
A sticker shortcut should, at a minimum include at least one !KEYWORDS entry in the form of Set <set number> where <set number> is the number of an official LEGO set in which that pattern appears. Only one set need be listed but more can be if desired.[/b]

[Chris Dee]

Based on feedback, I'd like to proposed a few minor changes to the Official Library Spec:

Clarification to Sticker Geometry:
- For stickers printed on a transparent medium, the type 1,3, or 4 lines that comprise the pattern shall be preceded by a 0 BFC NOCLIP statement and ended by a 0 BFC CLIP statement
- The pattern should be oriented such that the top of the pattern aligns with the +Z axis. Exceptions to this rule may be made on a case by case based.

Claification to Sticker Shortcuts:
- Require a !KEYWORDS entry with text similar to Pattern !KEYWORDS Entry:
A sticker shortcut should, at a minimum include at least one !KEYWORDS entry in the form of Set <set number> where <set number> is the number of an official LEGO set in which that pattern appears. Only one set need be listed but more can be if desired.[/b]

I approve, with a minor change to the second proposal: "... in which that pattern appears." --> "... in which that sticker appears."

[Magnus Forsberg]

Clarification to Sticker Geometry:
- For stickers printed on a transparent medium, the type 1,3, or 4 lines that comprise the pattern shall be preceded by a 0 BFC NOCLIP statement and ended by a 0 BFC CLIP statement

Why should we be forced to add a BFC NOCLIP in all stickers on a transparent medium?
I can see a need if the transparent sticker is intended to be on a tranparent part, like a glass window, but not if it's only used on solid coloured bricks. If the idea is that a user might want to have any sticker on a transparent part visible from behind, then why don't we have all patterns, on all parts, in a BFC NOCLIP? I don't see why a print on a transparent sticker should be treated differently than a printed brick.

- For stickers printed on a transparent medium, the type 1,3, or 4 lines that comprise the pattern shall be preceded by a 0 BFC NOCLIP statement and ended by a 0 BFC CLIP statement

Does this mean that the unprinted top surface is not allowed to be within the BFC NOCLIP section?

[Orion Pobursky]

If the idea is that a user might want to have any sticker on a transparent part visible from behind, then why don't we have all patterns, on all parts, in a BFC NOCLIP?

I'm not a fan of the proposed rule as I think it is too use specific but it was requested to I added it.

I also had the same thought as above. One of the major points of LDraw is that parts can be in any color. Maybe we should update parts as they are recycled. Or, better yet, just revise the BFC spec so that non-color 16 and 24 lines are excluded from BFC (actually, now that I think about it, that might be too broad).

I'm willing to live with a small quirk of rendering for the sake of simplicity.

[Willy Tschager]

Why should we be forced to add a BFC NOCLIP in all stickers on a transparent medium?

To be in sync with the specs for patterns on transparent parts and not have yet another exception of a general rule?

w.

[Willy Tschager]

Based on feedback, I'd like to proposed a few minor changes to the Official Library Spec:

Clarification to Sticker Geometry:
- For stickers printed on a transparent medium, the type 1,3, or 4 lines that comprise the pattern shall be preceded by a 0 BFC NOCLIP statement and ended by a 0 BFC CLIP statement
- The pattern should be oriented such that the top of the pattern aligns with the +Z axis. Exceptions to this rule may be made on a case by case based.

Claification to Sticker Shortcuts:
- Require a !KEYWORDS entry with text similar to Pattern !KEYWORDS Entry:
A sticker shortcut should, at a minimum include at least one !KEYWORDS entry in the form of Set <set number> where <set number> is the number of an official LEGO set in which that pattern appears. Only one set need be listed but more can be if desired.[/b]

Since there were no more comments for 3 months it should be added to the specs.

w.

[Travis Cobbs]

Since there were no more comments for 3 months it should be added to the specs.

w.

I issued a call for votes:

https://forums.ldraw.org/thread-24745.html

[Magnus Forsberg]

Why do this text say:

shall be preceded by a 0 BFC NOCLIP statement and ended by a 0 BFC CLIP statement

when the bfc page say:
https://www.ldraw.org/article/415

affects all lines following the CLIP/NOCLIP option, either until the end of the current file or another CLIP/NOCLIP option is encountered.

Why should the sticker be forced to be ended by a 0 BFC CLIP, when the other rule say end of file ?
Is it correct to end a file without a 0 BFC CLIP statement, if you have used a 0 BFC NOCLIP?

[Travis Cobbs]

Why should the sticker be forced to be ended by a 0 BFC CLIP, when the other rule say end of file ?
Is it correct to end a file without a 0 BFC CLIP statement, if you have used a 0 BFC NOCLIP?

If the pattern geometry is at the end of the file, then the 0 BFC CLIP is indeed not needed. It is however required if any non-pattern geometry follows the pattern geometry. So the wording should be updated to reflect that.

[Willy Tschager]

If the pattern geometry is at the end of the file, then the 0 BFC CLIP is indeed not needed. It is however required if any non-pattern geometry follows the pattern geometry. So the wording should be updated to reflect that.

Agreed. Could you please update the wording and also rework the requirement of a set number in the KEYWORDS for both stickers and sticker shortcuts?

w.

[Milan Vančura]

Sorry for a greenhorn question: does this solve the issue when transparent part with sticker looks as with no sticker when viewed from the back?
   

[Travis Cobbs]

Agreed. Could you please update the wording and also rework the requirement of a set number in the KEYWORDS for both stickers and sticker shortcuts?

w.

OK:

Clarification to Sticker Geometry:

For stickers printed on a transparent medium, the type 1, 3, and 4 lines that comprise the pattern shall be preceded by a 0 BFC NOCLIP statement. If there is non-pattern geometry after the pattern geometry, the pattern geometry shall be ended by a 0 BFC CLIP statement.

The sticker should be oriented such that the top of the pattern aligns with the +Z axis. Exceptions to this rule may be made on a case by case basis.

New Sticker !KEYWORDS Entry section immediately following the Sticker Description section:

All stickers and sticker shortcuts should, at a minimum, include at least one !KEYWORDS entry in the form of Set <set number> where <set number> is the number of an official LEGO set in which that sticker appears. Only one set need be listed, but more may be if desired.

To reiterate what I said elsewhere, I don't have a strong opinion about this. I did the call for votes because it seemed fine, and there appeared to be consensus before I did it. I don't really plan to argue the merits of the above changes, but I would greatly appreciate it if anyone has problems with them, that they speak up before another call for votes happens.

[Willy Tschager]

Look good to me.

w.

[Willy Tschager]

Sorry for a greenhorn question: does this solve the issue when transparent part with sticker looks as with no sticker when viewed from the back?

Read: https://forums.ldraw.org/thread-22204.html

w.

[Magnus Forsberg]

Sorry for a greenhorn question: does this solve the issue when transparent part with sticker looks as with no sticker when viewed from the back?

I would answer yes, and no, to your question.

The idea is to make the pattern on a transparent sticker visible from the back side, but all the parts in your image are printed, not stickered.
Please note that LDCad use the word "with sticker" in partbins, containing both "printed parts" and "parts with stickers".

Have a look at parts 60581d50 and 60581dm2.
   

The sticker with a thermometer is made following the new suggestion. The print is visible, mirrored, from the back side.
The print on the Moria stickers are not made using a BFC NOCLIP and are not visible.
Should the Moria stickers be visible from behind? They are not intended to be placed on a transparent panel. The panel is grey in this set.


Printed pattern on a opaque part is not requred to use the NOCLIP.
Printed pattern on a transparent part should use the NOCLIP.

Should prints on a transparent sticker be forced to use the NOCLIP?

btw.
The 2 panels in your image with a porthole is not correctly made. These are prints on a transparent panel and should have been made using a BCF NOCLIP.

[Roland Melkert]

Have a look at parts 60581d50 and 60581dm2.

Please note that LDCad renders the whole part dual sided when a noclip is present anywhere in the part's .dat tree.

And even then, it could not show the pattern from the backside as LDCad does no triangle level alpha sorting (only part level).

[Chris Dee]

OK:

Clarification to Sticker Geometry:

New Sticker !KEYWORDS Entry section immediately following the Sticker Description section:

To reiterate what I said elsewhere, I don't have a strong opinion about this. I did the call for votes because it seemed fine, and there appeared to be consensus before I did it. I don't really plan to argue the merits of the above changes, but I would greatly appreciate it if anyone has problems with them, that they speak up before another call for votes happens.

That's better now. Thanks.

[Milan Vančura]

Thank you for your answers, Willy and esp. Magnus. I've read the linked thread as well ("heavy" for me...) and Magnus' detailed answer here helped me to understand the topic more.

I know (and knew) about the difference between stickers and printed parts, although I do not know technical details. From my (user's) point of view, both types should work and both types should work regardless the "original part" LEGO released with either the print or sticker:

Stickers can be applied to any part of any color (of course)
In digital virtual world even printed parts may be used in any color, why not?

And, watching from the other side of the problem, it may be confusing even if you do not plan to use any printed or stickered part in new color: simply because you want to find the part with no print/sticker among all those parts - and watching "from behind" they look all the same. What is surprising and confusing, if the color is transparent. Look at the image I posted as an example.

BTW, I made a test: I added BFC NOCLIP to the definition of a print on this panel and it works in LDCAD both in the part bin and in the main model view:
   

[Willy Tschager]

That's better now. Thanks.

If all are on the same page I think it is safe to call for a vote.

w.

[Philippe Hurbain]

Clarification to Sticker Geometry:
- For stickers printed on a transparent medium, the type 1,3, or 4 lines that comprise the pattern shall be preceded by a 0 BFC NOCLIP statement and ended by a 0 BFC CLIP statement
- The pattern should be oriented such that the top of the pattern aligns with the +Z axis. Exceptions to this rule may be made on a case by case based.

Sorry to chime in late, but there is another point that could be clarified in BFC NOCLIP usage (for stickers or transparent parts): should color 16 (transparent section) of pattern be included in NOCLIP section or not? Case in point: https://www.ldraw.org/cgi-bin/ptdetail.c...973p08.dat
I initially thought that these transparent areas belonged to the pattern so should be placed in NOCLIP section, but this would cause inconsistencies with parts only partially patternable where most geometry is in a BFCed subpart, including areas next to patternable areas. So I now think that transparent surfaces should NOT be in NOCLIP section.

[Travis Cobbs]

Sorry to be anti-social, but I refuse to have anything further to do with this topic, other than voting on the current call for votes once the other standards board members do so, and voting on any future call for votes that they might bring.

[Orion Pobursky]

The part you used as an example is not a sticker. This rule is only for stickers, not regular parts. We can discuss expanding to all patterns but I'm starting to think that this whole ordeal wasn't worth the effort in the first place as the problem could be solved by the rendering software itself.

[Travis Cobbs]

OK:

Clarification to Sticker Geometry:

New Sticker !KEYWORDS Entry section immediately following the Sticker Description section:

To reiterate what I said elsewhere, I don't have a strong opinion about this. I did the call for votes because it seemed fine, and there appeared to be consensus before I did it. I don't really plan to argue the merits of the above changes, but I would greatly appreciate it if anyone has problems with them, that they speak up before another call for votes happens.

Changes have now been approved by the Standards committee. Can someone with the rights to do so please update the spec page?

[Orion Pobursky]

Changes have now been approved by the Standards committee. Can someone with the rights to do so please update the spec page?

Yes, I will

[Magnus Forsberg]

Changes have now been approved by the Standards committee.

Does this mean that all unofficial stickers that doesn't follow the new rules, have to be reworked and updated before a release.
I think there's some Admin certified stickers that maybe should be Hold voted now.
e.g.
https://www.ldraw.org/cgi-bin/ptscan.cgi...ope=header
https://www.ldraw.org/cgi-bin/ptscan.cgi...ope=header

[Willy Tschager]

Does this mean that all unofficial stickers that doesn't follow the new rules, have to be reworked and updated before a release.

Yes. No need to fix already official part (obviously they need a fix if uploaded to PT for an overhaul). Are you gonna hold or shall I?

w.

[Magnus Forsberg]

Are you gonna hold or shall I?

If you do the Hold voting, I could start fixing them, and then you can vote. But we need a third person to give the second vote.
I'm thinking we should fix these files, not chase the original author. After all, they have done nothing wrong. We have added more rules to follow.
How do you think we should handle this?

[Orion Pobursky]

If you do the Hold voting, I could start fixing them, and then you can vote. But we need a third person to give the second vote.
I'm thinking we should fix these files, not chase the original author. After all, they have done nothing wrong. We have added more rules to follow.
How do you think we should handle this?

Your suggestion is the path I'd take. I can be the third reviewer.

[Orion Pobursky]

Your suggestion is the path I'd take. I can be the third reviewer.

Once the changes are made, if you could post a list (nothing fancy, just part number is fine), I'll take action.

[Magnus Forsberg]

Once the changes are made, if you could post a list (nothing fancy, just part number is fine), I'll take action.

OK, here is the first batch:

6177949e  -  subfiles es01- es06, 6 files edited
6177949f  -  subfiles fs01- fs06, 6 files edited
6177949h  -  subfiles hs01- hs06, 6 files edited

4121723   -  k, n, o, p, and subfile s10,  5 files edited

6285381, a-h, ac01, cc01 and ec01, 11 files edited

[Max Martin Richter]

OK, here is the first batch:

6177949e  -  subfiles es01- es06, 6 files edited
6177949f  -  subfiles fs01- fs06, 6 files edited
6177949h  -  subfiles hs01- hs06, 6 files edited

4121723   -  k, n, o, p, and subfile s10,  5 files edited

6285381, a-h, ac01, cc01 and ec01, 11 files edited

I'll try to find some spare time as well in the next days.