Page 1 of 1


Posted: Sun Feb 03, 2008 3:43 pm
by robertw
Currently Stella doesn't print tabs on nets. Do people think it's important that tabs should be printed?

The FAQ has the following:
Q: Are tabs printed on nets?

A: No. Maybe someday, but I don't think it's too important. You just need to remember to leave tabs on edges as you cut out the nets.

There are three ways people might make models:
  1. No tabs - held together with either sticky tape, or even glue along the edges. Requires thicker card I think.
  2. Single tabs - at each edge to be connected, one face has a tab and the other does not, the tab being glued under the other face.
  3. Double tabs - both faces have tabs, which are glued to each other (leaving it "ribbed" inside, with tabs poking into the centre of the model).
Only with method (2) do you need to think about which edges have tabs and which don't.

I use mostly (3), with (2) occasionally for edges with very sharp angles, or where there's not much room in the net for cutting out two tabs.

With method (2), it's usually not hard to figure out where to leave tabs anyway. In particular, when the whole model is made of a single net, you just put tabs on every second edge around the outside silhouette of the net.

But I do recommend method (3) for most cases. In this case you just leave a tab on every edge. With tweezers it is easy to squish the tabs together (it's harder when gluing one tab under another face), and glued edges look more "symmetric" and precise. Also, when gluing a tab directly underneath another face, the face can often become wrinkly from the glue. It's also often somewhat easier to put that last piece in!

When I use a mix of (2) and (3), I try to glue all the (2)-style edges first, since they are easier to do near the start.
Thought I should get some users' opinions.

Tabs are actually kind of tricky to handle well in the general case. They may need to be made smaller in order to physically fit inside the finished model, or just to fit in with other tabs around the net. Otherwise I would have added them already, although in practice my model-making has worked just fine without them :)


Posted: Tue Feb 05, 2008 8:32 am
by oxenholme
Good topic!

I invariably use double tabbing now.

Sometimes (e.g. regular compound of Five Cubes with dodecahedral symmetry) two or three facets have collinear edges, so I will use a common tab to maintain linearity and increase the rigidity of the model.

I have to admit I find it a pain getting tabs the right shape. I like them to be as large as possible, but when one is double-tabbing on the Great Icosahedron for example the shape is critical if one is to avoid conflict and the resulting distortion.

Now where I really have a problem is on colouring. I'm colour blind and have considerable difficulty telling some shades apart. So when it's a stellated triacontahedron and I'm using fifteen different colours of card it can be quite fun getting things right!

Posted: Tue Feb 05, 2008 1:55 pm
by robertw
Here's a thought regarding your colour-blindness (which I just thought of, but is possibly common practice among those in the know!). You could look through coloured cellophane or similar to distinguish some colours. Eg look through something red and transparent and green paper will get dark while red paper will stay light. It's used in black & white photography to darken blue skies etc.

Did I just invent something useful? :D


Posted: Tue Feb 05, 2008 2:00 pm
by robertw
I agree sometimes it would be handy for Stella to figure out how big the tabs can be to still fit inside the model without conflict. Usually I just start bigger with the first net of a given type, and shave it down as required. Then I know for the next net, so it's not much of a problem.

And good tip about colinear edges. I used the same thing in my model of the Great Icosicosidodecahedron:


Posted: Sat Mar 01, 2008 8:45 pm
by guy
I usually use method (2) - one tab glued under the joining face. It's much the quickest way when it's done right. Also, it offers the most foolproof way to number the assembly sequence, matching Tab 1 to Edge 1, etc.

It makes a huge difference (especially to the beginner) how sensibly the tabs are placed and numbered. It is all too easy to place one or two so that the fiddly bits become impossible to get right. No tabs are far preferable to badly-placed tabs.
There is a simple algorithm that I have seen used, even recommended - put a tab on every alternate edge. Don't offer that!
When I design a tab layout, I think through the order of folding and gluing, making sure that the early bit has the tab and the latest face glues down on top of it. Except of course for the unpredictable situations when the other way works better. The later faces may have several tabs/edges with the same number to glue to, so I know to glue them all at once.

Anyway, I reckon that worthwhile tab design is an expert skill that would be a huge task to program - it would need to understand what fingers can and cannot reach, for a start.

Now, an interactive tab designer tool so I can point-and-click my way round the net and it will then fold up for me to check I haven't made any gross boo-boos. That would prompt me to upgrade!

Posted: Wed Apr 02, 2008 5:30 pm
by Nordehylop
I think that a tool to tell you how to fit tabs inside the model without any conflict would be very handy. For example, I had to trash a vertex part for my great icosahedron because the tabs were to big and they made the vertex point turn into more of a "vertex area".

On the other hand, I don't think an "interactive tab designer" is at all needed. For one thing, it seems that everyone does tabs differently, and it would require a great deal of work to program. The capability to print out nets for a couple trillion polyhedra or so is good enough for me :D.

Posted: Mon Apr 07, 2008 9:35 pm
by guy
Well, I guess one man's tool to tell you how to fit tabs inside is another man's interactive tab designer.

What I envisage is the ability to move tabs around and check whether they need a bit of trimming or similar.

For example it might:
* keep track of which edge pairs with which, and insist that just one of these has a tab.
* Offer a default of alternating tabs or sequential tabs or whatever (maybe even a choice) as a starting point.
* Check for slender vertices and trim the end of the tab accordingly.
* Allow you to do things like right-click > swap tab.
* Offer an assembly sequence, with printed numbers that you can also edit.

Posted: Mon May 04, 2009 2:21 pm
by the_tool_man
I've only been using Stella for a couple of weeks, but I find myself wishing for printed tabs, as well. For me, the difficulty lies in predicting the taper angles at the tab ends so that the tabs at a given vertex don't interfere with each other. I have managed, so far, using trial and error, but as the models get more complex, I see this being more of an issue. I only use double-tabs, and have assumed so in the following description.

In operation, the user could specify the tab width, either as a dimension or as a percentage of the polyhedron scale. The taper angle of each end of each tab could be figured from the geometry of it's ajacent vertex. This would result in trapezoidal tabs on each edge (or triangular if the tapers are shallow enough to truncate the tab). You could add a checkbox to your print preview form to print tabs with the nets.

As far as using single tabs, it would seem that the computational overhead of keeping track of which net edges join up would make automatic single-tabbing difficult. However, you could have the option to select individual tabs for removal. Or, the user could simply trim off the tabs that are not needed.


Posted: Tue May 26, 2009 2:29 pm
by robertw
Hi, sorry for this reply to take so long, but just a couple of comments.

I agree, tabs would be handy. I've already got it somewhere on my to-do list, including the idea of calculating the angle to taper out the tabs at each end. There'd be some maximum angle, but a lesser angle as you say when called for. And the taper angle will be different for people who want to use single-tabs.

It's not quite a simple matter of adding tabs to the nets as they stand though. Tabs may intersect with other parts of a net, or intersect with each other. So there's the choice of whether to trim them to avoid this, or rearrange the parts of the net to avoid the problem (not necessarily possible if a single face has edges that come close together, causing their tabs to interfere).

It's all quite complicated, or at least time-consuming, to do right, so I haven't ever got around to it. I've also managed to make all my paper models just fine without printed tabs. Sometimes there's some trial and error on the first of a particular type of net, to see how narrow the angle has to be on the tab, so yeah it would be convenient for tab angles to be calculated and printed, but it's not vital. Someday I'll do it.


Posted: Wed Jun 03, 2009 8:41 pm
by guy
One thing that should be straight forward would be to mark the angles at the ends of tabs. For two-tab people, that's all they'd really need. For us one-tab people, there's always tomorrow.