Press "Enter" to skip to content

Why Square Corners?

Geometry is important in pattern making.  I know I go off about that a lot, but it’s true, and it’s why I’m obsessive about squaring corners.  Here’s the thing: if your corners aren’t square, then your seams won’t meet squarely.  When your seams don’t meet squarely, you get weird looking edges on your sewn piece.

Here are examples of the two most common mistakes I see from students truing up drafts.  Check the waistline action:

1. Deciding that the longer part of the line to be trued must be correct (red line), and ignoring the inconvenient other part (black line).
Mistake #1. Deciding that the longer part of the line to be trued must be correct (red line), and ignoring the inconvenient other part (black line).
2.  Wanting to follow the curve of the ruler/curve (red) even with it departs from the edges of the draft (black).  I call this condition the Joy of Curve.
Mistake #2. Wanting to follow the edge of the ruler/curve (red) even with it departs from the edges of the draft (black). I call this condition the Joy of Curve when my students come down with it.

And here is what happens: when seams don’t meet squarely, you get peaks and valleys.

This is what happens with you have an obtuse (larger than square) angle going into the seam.  That's what happened in our first example.  Notice that the waistline is not straight - it angles up at the center.
This is what happens with you have an obtuse (larger than square) angle going into the seam. Notice that the waistline is not straight – it angles up at the center.
This is what happens when you have an acute (less than square) angle, like our second example.  Notice how the waistline is really not straight.  That's the Joy of Curve for you: both sides of the waist have been extended into odd points.
This is what happens when you have an acute (less than square) angle – you get pokey bits. Notice how the waistline is really not straight. That’s the Joy of Curve for you: both sides of the waist have been extended into odd points.

How many of us have done this with bodice straps?  *sheepishly-raises-hand*  I was young!  I was new at this!

This is annoying at a waist or a strap, but imagine it happening at the side seam/armscye area.  Try setting that sleeve!  It’s not awesome, people.  :(  To prevent it, make sure that the corners of your draft are square.  It’s easy to check with a gridded ruler.  The whole line doesn’t have to be square, as you can see in our waistline examples – it’s better for a straighter line to be square for some length (like our waistline), but a curve (like an armscye)  just needs to be square for a moment at the corner.

For my fellow geometry nerds, there’s a caveat: technically, as long as the angles of the two pieces in the seam add up to 180 degrees, you’re fine.  Go nuts, dare to be festive, draft dangerously!  If you’re not sure what that means, of you can’t quite remember why 180 is important in geometry, then stick with squares.  You can’t go wrong when your seams meet squarely, and also you won’t need a protractor; just your gridded ruler.  ;)

So, can you just guess how insane it makes me that the 1/2 scale slopers in the text we use aren’t squared at the side seam and center front neck?  Nnnnngggggrrrrr!!!!!  *falls-on-hands-and-knees-pounding-the-earth-shouting-why-god-why?!?*  (Yes, it really is that dramatic of a thing for me.)  I keep reflexively squaring them.  I try to stop myself, especially when I’m doing a demo, but it so hurts my OCD.  :/



  1. Deborah
    Deborah February 10, 2014


    • missa
      missa February 10, 2014

      :) Word! The mini-bodices helped this time.

  2. […] Step 3: Drafting a Pattern. Using a ruler and french curve, I cleaned up the somewhat jagged traced lines. I just used a regular ruler for the straight side seams, and the curve for the arms and neck. I’m not going to get into the french curve here today, because I think I need to use it more before I can explain it very well! The tricks I used in drafting are really a post in of themselves.  I found this great explanation about squaring corners on Sempstress. […]

  3. Hi,
    I wish I’d taken pattern drafting in the spring. ( I took sewing I and now neither sewing II nor pattern drafting I are given in the summer! That foiled my master plan! Argh!! )

    I’ve been going ahead and trying to write a script to create some drafts anyway (and sewn a few up This is periodiaclly broken or working. Right now it’s “wonky” because I’m fiddling with trying different methods to deal with ABC cup alterations. I started with a block from winnefred Aldrich’s book, which I checked out of the COD library and have been reading two different methods for slashing and spreading for the bust cup. I’m now a bit torn over which is best and so interupted coding. (not because of the math, but because I think armholes get ‘weird’ looking if a take a cut from the mid-armhole — and I don’t know if that ‘weirdness’ is a bad thing or ok. This is where lack of experience fitting to a number of bodies is my problem. But I’m not going to ask you that because it’s not the topic of your post! )

    Anyway, I have a question on squaring of the block. In your figure of “what happens with you have an obtuse” to the left you show the block. I can see the red diagonal line at the waist intersecting with the center front (black vertical parallel to length wise grain.) That forms the obtuse angle. I assume the muslin bodice to the left was cut using that red diagonal line? (Not the black trace below which becomes horizontal as it approaches the center front? Right?

    Anyway, I can see the waistline angles up at the center. But it appears the length looks “ok”. So assuming the answer to the above is “yes, it’s cut from that red diagonal line” my question is: wouldn’t cutting at the block trace below make the center front too long? (Basically, if I zoom in at the bottom of the block , I see a triangle with a red hypoteneuse and two black legs. The vertical part of the leg adds extra length. It seems to me preferable to come up with a trace that is horizontal, doesn’t add this extra bit length. Sort of like the skanky purple trace here (×300.png). Is my sense right?

  4. missa
    missa June 5, 2014

    Hi, Lucia – Wowza! Ok, so the short answer is “yes”. ;)
    The longer answer is: Sweet! Are you really totally working on a php-based flat system? Because, like, that would give be something to live for. I’ve wanted to work on something like that for years, but I did not even know where to start. Math-based programming is not so much my strong suit. Any chance you’d be interested in a collaborator?
    As far as the red line issue, if you take a look at the bodice that angles up you’ll see that the center front is actually too short – even with the seam allowances on, it’s barely covering the waist tape on the stand. (I admit the seam allowances really help to make this unclear – I should have put in a stitch line indicating the finished bottom of the pieces. Apologies!) That means it’s fallen at least a half inch (at half scale) off course. The black line below is actually the correct line for the bodice to be cut on to reach the waist when finished. It comes directly from the block I was altering – I just tossed the dart over to the side seam to make a point about squaring and truing up after dart manipulations for my class. So basically, your sense is good but my picture was misleading. I think I’ve still got those little mockups in the office if you want to actually throw them on a half scale at the college and see how all everything lines up.
    Oh, and not that you asked, but… The easiest way to deal with cup size alterations is (imho) to work off a princess line block. Presto-change-o, now you can alter the bust without impacting the armscye or upsetting the CF line or anything else wonky. ;)
    Hope that helps!
    ps – sorry the summer scheduling foiled your master plan. :( I have a T/TH class over the summer, but I’d be happy to get together and help with any pattern stuff when I’m not in class. Feel free to drop me an email.
    pps – what’s up with the complementary angle set in your side seam in dynamicPatterns pic?

  5. missa
    missa June 5, 2014

    erg – I should also say that the location of the curve is determined by the dart that was closed, and it really can’t be moved far because, the human body. It goes back to the actual function of darts in creating curves on the body, and the fact that the human form lives neatly between cartesian geometry and a polar grid on acid – the lines that are straight horizontals on us when viewed would really be curves if we were laid out flat. This is why we talk about horizontal balance lines in patterns, and why they’re so linked to layout when cutting: you have to pick something on the body to represent a truly straight horizontal line, or you will never ever be able to get fabric to hang right since woven fabric is (theoretically) completely orthogonal in nature. It’s why circle skirts distort over time and off-grain pieces ripple and pull – you’re working against the fabric. So we use horizontal balance to establish grain.

  6. Yes. I’m like totally working on that! :) My main difficulties are knowing what would really, really work. (That’s why I want to take the pattern drafting class so I’m doing better than just self teaching. I should have just bit the bullet and taken both classes in the spring. Ellen was great in sewing I, but…well, I out foxed myself there. I tried to get COD to get your class to be a hybrid this summer… so I could at least get my sewing better.. but no dice! )

    On the math: I have a phd in mechanical engineering. The php is self taught. To some extent, my main challenges are still: not knowing completey how to “organize” things due to not entirely understanding pattern drafting. The result is I’ll code something in a way that seems “easy for now”, but only later do I learn which things are done over and over and over again and which things are best done one off. So, the code is a bit spaghettish.

    So: once again, I want to take pattern drafting (and more sewing so I know whether things I drafted have ‘issues’ do to my poor sewing or whether it’s my crummy drafting!)

    I’d love a collaborator!! Love, love. Most people I explain this too are just puzzled. (Other than my husband, who is also an engineer.)

    I did see your name was switched with teaching construction I. I was curious, googled and found your blog! I could easily stop by and show you some stuff and explain ideas. (Like, for example: the armhole curves are ‘bezier curves’. I have a subroutine to make bezier curves with any number of points one likes– though generally 3 or 4 is appropriate. I’ll explain how those work if you don’t know.

    Oh. php has drawing modules. I just use those.

    I admit I thought the center front might be too short– on the other hand, that depended a bit on whether the muslin included the seam allowances.

    By the complementery angle on the side seam– do you mean the weird wonky, it goes out and then in again? That’s “Thought I would try a new way to insert the extra spread for the cuppage and didn’t finish the job by coding in the full slash and spread operation.” I could show you the equivalent on paper if I stop by your office or Tuesday or something. That has to go away. But to do it right, I guess talking to you about the ‘better’ or worse way to deal with the bust spread would help. (The subroutine contains 3 tries 2 of which I commented out between /* and */!)

  7. Ok… reading this, it seems to me what this mostly means is: We may need to discuss the best *approach* to ultimately creating a ‘dynamic pattern’ that permits a person to customize the idea we envision for the outfit. (At least maybe that’s how I’m interpreting the ‘picking something to be the straight horizontal line’. My sense is I don’t grasp some of this enough to make some proper coding decisions. Hence: wanted to take pattern drafting!)

    My thought had been:

    1) Create a general block with some degree of ‘standardization’ (as in Burda, Simplicity etc.) but with a bit more flexibility. Changes for bust would be done here.
    2) Using those code in rotations and new pieces to make it a “whole pattern” for something like a blouse.

    But maybe it makes more sense to spread for bust after (2)? Am I grasping this sort of right? Or wrong?

    Anyway, the ‘master plan’ is to have a few patterns up that let people customize what they want to some extent. For example: the slacks I made for my husband here: I used the ‘standard’ block, coded all the pieces, and then saw he has a flat butt. So I found a discussion of slashing and spreading for ‘flat butt’ coded that in– and now the pants fit better without the droopy drawers thing going on.

    If I felt confident with those slacks, I could have pulldown menus for standard choices and some degree of customization and people could create their own pattern is more-or-less standarish sizes from size ranges like 26″ to 45″ waist picking “flat butt” “normal” or “generous butt” and also “short”, “medium”, “long”. (Greater customization would also be possible. ) Because it’s code it’s relatively easy to do this– within some limit. ( For example: my first fiddling with spreading a block for the bust using Aldrichs block “broke” the algorithm at the armhole when I went for DD cup. That’s when I started hunting other ideas.)

  8. missa
    missa June 5, 2014

    Hooray for collaboration! PhD in ME? Yeah. Ok, then. Your maths are stronger than mine. :) I’m better at doing up bezier curves in the form of string art than in math – never really got past pre-calc. So, funny story, but I’ve spent many the year programming. I ain’t afraid a’ no spaghetti. It’s mostly a hobby thing these days (php/sql/jQuery for WP), but I may be able to help with code organization. I’ve been working on a couple different new ways to do up blocks – the mathematical one might be useful to you. I’m still knocking off a couple rough edges, but, well, you know how nothing is ever entirely done? One of the troubles with pattern drafting is how the human body is capable of taking on quite so many shapes and sizes. I’m not sure any one single method words for all of them, but if it’s out there, I would like to find it.
    Yes, often talking about pattern drafting is the fastest way to put people to sleep or loose them. It makes teaching it a challenge. ;)
    Somehow I think I knew php had drawing modules, but I don’t think I’ve ever looked into them. I know it’s got pdf generators – used those. But never got to the drawing.
    Yes, that’s the weird but I was talking about. It will create a crooked side seam.
    I would ask if the code is on github, but then if you said yes I would immediately go dashing off to read it, which would not help me finish the project that has to be done by Monday. Or get ready for class on Tuesday. Le sigh.

  9. missa
    missa June 5, 2014

    lol – as we both respond at the same time!
    How’s about I show you how to draft to custom measurements? It’s going to be easier to code to start with. THEN it’s an easy thing, if you want, to define semi-standard options for the “But I do not wish to take that many measurements” crowd, because you just run the measurement-block-generator with mostly pre-defined numbers.
    Block manipulation via computer is going to be harder than creating a draft generator.
    Ah! That explains the angle of the upper back pant! Makes sense. (I was exploring your blog. I think WP will do everything you’re talking about setting up. I can help with that.)
    A DD cup will actually break most flat pattern manipulation methods. It’s just a thing.

  10. Some of the math I know is utterly useless for pattern drafting. But some is useful. My area was ‘fluid mechanics’– so pretty useless to pattern drafting. (Oh.. and of course taking a class on “Tensor Manifolds and Applications sounds like a big help. ;) Well… actually, that’s not quite as ironic is it likely sounds. Your slopers are two-dimensional manifolds and the lines you are thinking of curves are 1-D manifolds. And when you ‘walk the armscye’ to find the length, you are measuring the length of the curve! And when you are thinking of how the sloper looks on paper vs. on the body… well… you are changing coordinate systems. All without a jacobean in site! But actually taking the class would be useless for pattern drafting. )

    I have to credit the idea of coding in the bezier curves to my husband. I was grousing at breakfast, “I have to do curves. Spline interpolations… (grump, grump, grump).” He said– look up bezier curves. Oddly, I had done bezier curves in 5th grade art class, but never in math. He’s never used them at all, but knew they got used by people working on auto-bodies and stuff like that. Turns out the math is are very easy to code. It’s also very easy to take the derivative, so it’s easy to figure out how to impose (a) points the curve must go through (end points) and (b) the slope at the end points. So, for example, if I were coding the purple line I put in by hand on your draft, it would have 4 points. The first and last would be the end points. The interior ones would be to impose the slopes we like. Then I tell the module to calculate “N” points in between, and draw lines between those N. (I can make N whatever I like, but generally 10 seems to be smooth enough for every curve I’ve tried.)

    On the teaching is a challenge: I knew one of the women in your spring class. She thought the class was hard. I had to bite my tongue and quietly think “It’s not going to be hard for me.” In fact… I suspect for much of the stuff, I’d be able to explain even though I haven’t taken the class. (I still need the class because just reading and knowing some geometry leaves huge gaps. Like…what’s ‘balance’? and how do I know if I’ve achieved it? Rotations? Arithmetic? Algebra? Trig? Geometry? That’s not going to be a problem for me. On the other hand, sense or proportion? What actually looks good? Not sure my talents lie in that area…. )

    I’ve got the pdf generators going. I’ve printed these out, taped them together and made pants and simple tops. But the patterns aren’t user friendly.

    On the custom measurements: I checked out Joseph-Armstrong from the library. Coding that sloper would be do-able. It’s not any harder than coding the simpler block in Aldrich– just more measurements. In some ways, the ‘from many measurements’ may turn out easier– because Aldrich contains numerous arbitrary constants which replace the need for numerous measurements, but also clearly assume some particular figure type. (Or more specifically: make a larger number of assumptions that the ‘from custom’.)

    I also found recent anthropomorphic measurements at a site selling mannequins. (I need have the library order the full ASTM circular!) So it’s easy to code in ‘standards’. But going forward, I know I want some of the non-standard features to be changeable in a ‘user friendly way’. (For example: most home sewers aren’t going to easily measure things like ‘Shoulder slope’ etc.

    On the DD breaking most flat pattern manipulation methods: Ahh!!! Ok. That makes sense then. I was looking at various ways to incorporate that,
    1) The first way– just increasing the neck to bust dart in Aldrich totally screwed up the armhole (in a hilarious way).
    2) Then I looked up Joseph-Armstrong–Slash from mid arm to bust dart. Armhole… not as screwed up, but had such a ridiculously tight hair-pin curve, that I thought… hmm… That’s got to be uncomfortable.
    3) Thought up on my own: What if I Slash from the top of the armhole/shoulder point…. Reduced the ‘hair-pin’ curve somewhat, but I haven’t read that suggestion anywhere. So, I lacked confidence.
    4) Google: Found U new mexico. Showed slower with side and waist darts. Slash from midshoulder to bust point, slash through dart. That’s what’s “half” coded (even though it doesn’t look like it because I didn’t code straigtening out the bottom).

    This code is not on github. I’m not really planning to share widely. It’s just been “me”. So, if I were sharing with someone to look, I’d just .htaccess password protect and grant access that way. That said, right now, there is no real ‘security’– it’s just stored on a little trafficked site! It’s all spaghetti anyway though– or at least 1/2 spaghetti.

    This is getting long. I think we need to sit down with books, paper, coffee and talk about what the “vision” for “target functionality’ is– and then think about the organization and what is the best approach. I live in Lisle, and drive to Arden Courts 3 blocks south of COD frequently to see my mother-in-law who has Alzheimers. I’m retired– but with the plan to get this going. So my time table is super flexible. (I guess I should just go to the COD page and phone!)

  11. Finding your email at COD’s page is too challenging. I’m following you on twitter. If you follow me, we can exchange emails by direct message! (As I follow you, I think you can already direct message me.)

  12. missa
    missa June 6, 2014

    My email isn’t on the CoD page – they only list full-timers. I sent you info via twitter.
    For the record, when I took flat pattern I’d already been patterning and sewing for years – entirely self taught, for historical. It was still challenging for me. A large part of that was learning to get out of my head and just try things on paper. There’s always that temptation to think that you can totally work it through in your head so that you won’t ever risk making a mistake and having to redo. I mean, that and the draft in the Armstrong book totally falls apart on a curvy girl. You’ll find that every draft has some biases built in about what figure it’s expecting. A lot of them are biased towards an A/B cup. Moulage deals much better with a variety of curves in front, but the amount of curve it accommodates in back is basically hard coded into the methodology. Go back to the 17th century, and most of the process was hardcoded into the methodology based on “ideal” measurements. It’s actually fascinating how it’s all evolved.
    Do you mean balance in terms of pattern making, or balance in terms of visual design?
    Option 4 (midshoulder slash) is creating a princess line as part of its process. That’s the most sensible way to do what you want to do. Never screw with a working armscye if you don’t have to. Option 3 will work for smaller alterations – you won’t change the shoulder slope, but you’re going to pull the top of the armscye off square (changing the relationship of the armscye and shoulder line), so be aware.
    Github isn’t just for sharing – it’s the great mighty undo in the sky. ;) I don’t use it like I should.
    Definitely time for the sitting down with the paper and the planning. My T/TH ends at 2.50 – does meeting at the college work for you?

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.