View Full Version : Curved Hips and Flared Eaves ?

Joe Wood
01-22-2005, 01:12 PM
I’m having a little trouble designing a curved hip and flared eaves on this Dutch Hip roof I’ve drawn.

Joe Wood
01-22-2005, 01:13 PM
I’m looking to do something kind of like pic A.

Joe Wood
01-22-2005, 01:15 PM
You can see in pic 4 what I’m trying to do, adding curved cleats to the top of the hip and those two jacks. Well, I’m having a hard time drawing them accurately, and probably will have as hard a time cutting the actual pieces accurately too, so that everything planes in correctly.

I was wondering if anyone has used a different technique for getting the curve and flare.

I’ve researched it in quite a few books of traditional Japanese and Chinese Roof Construction and, it’s too complicated doing it that way!

Joe Wood
01-28-2005, 08:37 AM
So, doesn't Anyone have any ideas for me ? I'm kind of stumped on this one, and can't probably see the forest for the trees ..

Tim Uhler
01-28-2005, 08:45 AM

I'm about to take off for work, but when I get home I'll respond. I have some suggestions similar to Joe Carola's at breaktime.

I would love to see pics of this in progress :-)

01-28-2005, 06:00 PM
Joe -
I'm not sure what kind of advice you are looking for --

Is it how to layout the Hips ? If so the hips are ellipses formed by the run of the hip....
they can be laid out by projection or by Rise & run of the 1/4 ellipse --

Am I on the right track here ? Or are we talking structural ideas ?

Mike S

Joe Wood
01-28-2005, 07:00 PM
Well, I'm asking mostly how to build it, the compound-curved part anyway. You see where I'm headed .. adding cletes to each hip and jacks, then bending 1x4 T&G down over it.

Actually, I'm also asking for a way to figure it out I guess, so I could cut, and lay down, tapered cletes. I could even make-up a fairing block, and fair everything in.

Or, is there some other way ?

Here is how you're supposed to buildem, but after studying Books on Chinese-Buddhist Roofs now for 2 months, I can't Even figure out how you're supposed to ! I'm looking for a Western Way, or a simplified Asian Way (preferebly) ..

01-28-2005, 10:36 PM
Joe -
my response it a bit complicated --- However turn the whole thing upside down.... as if building a groin vault ...

the hips are ellipses .... You may want to project the curves on a plan view ... Or
layout the ellipse using the run of the hip --- This is a subject that needs pictures & graphics ... I really am handicapped in that dept.

I am doing a 1 hour workshop on establishing ellipses at JLC Live in 2005 --- It's more of a hands on demo ....

I think Joe fusco could probably generate the graphics to explain this ... If he's still lurking around the forum ....

I might recommend George Collins - Curvature & Circular work - a book written in the 1800's & updated to modern english - Craftsman books might carry it -- But I'm sure it's available on Amazon

After doing much research on Curvature work of our predecessors, I realize that they plotted everything ... Math was not a major factor in their layout work. Thats why the Framing square is such a basic tool ...
The pieces you want to make can be easily plotted in plan view in 100% scale .... But for me to explain it online ??

Give me some Idea of the radius of the 'common tail' you want to cut -- If it's a regular hip - I can tell you the plot of the ellipse hip - Of course the compound flair as it approaches the hip will change the dimension.

Once you SEE the relationship it will all come together ...

Have I made any sense ?

Mike S


Joe Wood
01-28-2005, 11:14 PM
Geeze Mike, I was just trying to layout a radius for you, but the problem is, each jack will have a different radius, and it's not one radius that will define each jacks' curve, but more like one gradual one and one tight one. The hip will have an even more complex curve I'd think.

You sound like you're pretty on to the concept though ! Wish I could read your mind.

01-29-2005, 01:26 AM

I know from your previous posts that you are a bit fanatical about spacing and symetry on your hip structures. I hate to risk bringing this thread down a notch...but we would probably forget the math on the jack rafters and just establish a pleasing sweep on the hip, cut 4 scabs (for the 4 hips), tack in place and lay a straight edge on the hip and first jack (with no scab) and measure the space every 6-8" to get the "plane".

Especially with your softwoods, you can belt sand any high spots before installing the roof deck.

You are not trying to build a swiss watch. 10-12 feet off the ground and roofed with shakes, no one will know you did not figure each ellipse (smile).


01-29-2005, 05:44 AM

Are you taking about a curve or a true elipse They create two entirely different shapes which require two different methods of working them out.

Your problem is two fold, the mathematics and the construction. You can build formers and laminate all the members. You can get exceeding deep joist material and cut the shape or you can stack splice the timber and then cut to shape.

A while back there was topic on how to calculate elipses which I contributed to.

If I were you I would seriously consider a full scale layout before you start cutting any timber. A full scale layout will also provide you with the dimensions the timber has to be to achieve the correct result. I"ve built numerous convex and concave hip ends for verandahs but never the curved jacks.

Have you ever done any geometrical stair work? If so it's a simple case of creating the curve on the edge as opposed to the face.

The roof shape you call a Dutch Hip is in fact a Gambrel roof.


Paul Conrad
01-29-2005, 06:51 AM
Maybe in the Land Down Under that is a Gambrel roof, but here in the USA a Gambrel roof is a roof of two or more pitches used to give more space directly under the roof, often but not exclusivly used for barns. (It makes a good hay loft.) In the picture below, taken from the Webster dictionary, we see four roofs. Here is the caption from that picture:

roof 1a(1): 1 gambrel, 2 mansard, 3 hip, 4 lean-to

Also as a side note, I spent a year in Korea and enjoyed touring the Buddhist temples, but I also could not understand why they built them the way they did. They looked to be a little over engineered. But they were very beautiful examples of building/woodworking. I would never have the patience to do the painting on them. Good luck with your project, Joe.

01-29-2005, 08:54 AM

A 17th Century French architect called Francois MANSART (not Mansard some one even corrupted the poor buggers name) designed a roof system which consisted of a modified Queen Post truss with a King Post Truss sitting on top. The reason behind this was to create liveable space from within the roof system. This in effect created a 3 storey house more accommodation on same footprint At the change of pitch a horizontal beam was used called a "curb"(curb meaning restrain or hold) beam. The curb beam sat on top of the steeper rafters and the lower pitched rafters sat on top of the curb beam. The first roofs had gable ends. This design became very popular in continental Europe especially Paris. It helped to accommodate more people in a smaller area (sound familiar) This basic design then evolved, the original design was used in the crowded street of Paris. Large houses built out in the country where there was more open land started to have the ends with a double pitch. They looked far more fancy and flash (there were posers in those days too). To differentiate between the two the word Curb roof came to mean a double pitch on all sides. The French also gave the curb roof a fancy name which escapes me at this time.
The Dutch also took up the basic design but they built masonry gable ends like a parapet wall mostly with curves which extended past the roof line. That is where the term Dutch Gable originated. That style is sometimes called Dutch Colonial or Cape Colonial because that style of roof was favoured by the Dutch colonists in South Africa, originally called Cape Colony. The Dutch also built wind mills with a mansard style roof and the Poms pick up that idea. Now when the Europeans started to migrate to America they brought their knowledge and then started to adapt European design to local conditions. Many terms then started to be labelled and attached to different nationalities. In the North East of the US many Dutch people were the first settlers (remember New York was originally called New Amsterdam). It was the Dutch that originally built the Mansart (Mansard) roof in America but they then started to build entirely in timber as opposed to masonry and clay tile as in Europe. That is where the name Dutch Barn came from which is a barn with more than one storey in the roof style of the original Mansard roof.
It wasn't until the mid 1850's that some American decided to call the Mansard roof a Gambrel Roof. Now the Poms built their roofs differently to gain liveable roof space they built a collar tie roof and single pitch on the sides. To gain more space at the roof ends that started with a masonry gable end and about two thirds up they changed the roof to a tile hip end. That design was modified to be built totally out of timber. The correct and original term is Jerkin Head. Once more it got corrupted in America to a Kentish Gable or Tudor ( because that design originated in the late Tudor era and many such house were built in the southern English county of Kent.

The roof end which starts in a hip and finish in a gable is gambrel roof. The word gambrel came about because the outline resembled the shape of a horses hind leg.The small gable end was to allow ventilation in the roof space, an idea which originated in Java and Sumatra which was then also known as the Dutch East Indies. So that design of roof came to America by the Dutch. You call the same roof as a Dutch Hip and in OZ many people quite incorrectly call it a Dutch Gable If you want to give it a more correct name I guess you should call it a Java or Javanese roof because that is where the style (originally as a thatched roof) originated.

A lean to roof is also called a pent roof.

There are many types and styles of homes which are given corrupted fancy names by real estate marketeers, purely as a marketing ploy. Misdescription is practiced in OZ just as well and often as in the US.

Some of the corrupted names that come to mind are Colonial, Tudor, Bungalow, Californian Bungalow, Spanish, Mexican, Georgian, Victorian, Regency, Gothic, Cotswold, Queenslander, Mediterranean, Cottage, Tuscan, Provincial, Chalet. Sound familiar, and if you have seen any of the real styles in their country or region of origin you would realise what a load of BS is spoken about styles and types of houses


01-29-2005, 12:18 PM
Joe -
I think The thread to follow is understanding the Ellipse , It seems to me that it's construction is a form of the ellipse Or possibly a parabola. But understanding the principle of plotting out an ellipse from a basic radius might send you off in the right direction ... and you could probably carve out a decent represntation of it. - If it were a radiused tail that intersected a hip without the upswing- It most definately is a 1/4 ellipse. However as I think through the proccess it might be a series of ellipses that change major & minor axis as they converge in the corner.

It appears to me that the form of the roof you are trying to build is more in the Korean tradition than Japanese with the upswing of the roof hip. It is incredible carpentry & I wish I was young enough to study & master it.

I would love to follow this as you figure it out . Take a LOT of pictures.

I will be doing a short 1 hour Presentation at JLC Live this year on the relationship of ellipses to circles if you happen to be in one of the show areas ...


Mike Sloggatt

01-29-2005, 03:58 PM

If the hip rafter is in fact an elipse from the ridge to the end of the rafter tail, then all rafters, commons and jacks will also have to be an elipse. The hip rafter elipse will be the reference for all the rafters. The elipse for the common rafter will be a constant but but the elipses for the jack rafters will be all be different to come into plane with the hip elipse. The other issue is whether the hips, jacks and commons are to be cut parallel or is only the top edge eliptical. The roof pitch angle and the length of the rafters will also have an effect on the size (depth) of the rafters.

The more you think about it the more things come to mind that you have to consider. I reckon you could spend more time doing the calculations and setting out than pitching it.

I don't believe it is a project for the unknowing without a great deal of on site guidance. If the roof timbers are exposed then there is no room for error.

All those Oriental roof framers are master craftsmen who have been indentured and well and truly taught and strictly supervised over many many years. How many western roof framers can say that they were taught and learnt under those conditions.


01-30-2005, 12:03 AM
If the hip rafter is in fact an elipse from the ridge to the end of the rafter tail, then all rafters, commons and jacks will also have to be an elipse.

I think you missed something. Better take another look at the shots Joe posted. This thing is not that complex.

01-30-2005, 01:23 AM
Henry P,

Read and look at the pictures of the first 8 posts and tell us all what is really required because it is getting all very confusing. Mike and I must be barking up the wrong tree. There is one graphic where the ridge is curved or is an elipse. That would require even more working out whether by calculation or full scale set out.

In my post #11 I did ask the question whether it was a true elipse or a curve which is required.


01-30-2005, 02:05 AM

With this type of architecture, the hips are lifting. The common rafters and bulk of the eave are left alone. To create a graduated sweep to each hip, the adjacent 2-3 jacks (depending on the height of the "lift") will need scabs on top of the rafters as well. The common rafters and bulk of the jacks will be left as is (original dimension).

I think the confusion has to do with a true Mansard in which the entire eave would contain the sweep characteristic.

The only thing I might add to my original post in this thread would be to possibly "scarf" in these scabs above the affected rafters (including hips) possibly to a 1" depth so the taper scarf pieces do not feather to "zero"...this might look like the scarf was more "planned" in any case.

Joe must be on a long weekend spending some of 16K he is going to make in the next 2 weeks (sorry if this is going over anyone's head...you will need to find his thread in Bob's Estimating forum).

Earth to Joe Wood...come in Joe...

Almost said "Houston to Joe Wood"...but feared to create the arena for another pissing match between Dick and Allan (smile).


01-30-2005, 03:23 AM
Henry P,

I cannot get the quote function to work. I highlighted the sentence left clicked on the quote icon and the whole of your last post became the quote. What did I do wrong I must be rusty?

You said " I think the confusion is with the true Mansard in which the entire eave would contain the sweep characteristic"

You must be joking or have you been on the turps again, smile.

A true Mansard is a combination of a Kingpost and a Queenpost truss, no curves at all. Are you wanting to debate my post to Paul.


01-30-2005, 03:33 AM
Thanks Chippy. I stand corrected.
I meant to say "Dutch hip". (smile)

01-30-2005, 04:03 AM
Henry P,

No you didn't, you really meant a Gambrel Roof"


01-30-2005, 01:50 PM
Chippy - Just a clarification on one small point -
On a simple curve roof - The commons and jacks are chords of the same circle - not ellipses -- The radius does not change - the descending jacks are just smaller chords of the same circle. The Hip is a true section of an ellipse. typically less than a 1/4 ellipse on the type of roofs I might be asked to frame.

JOE Wood -
Here is an excersise for you :
take a piece of round 2" pvc pipe . Cut it at 90* -
Look at the cross section - it is a circle
cut the same pipe at 45* Now look at the cross section... It is an ellipse.
Technically the round 2" pipe has a major axis of 2" and a Minor axis of 2" that is a circle,,,,
Now the 45* cut you changed the Major axis - BUT not the minor axis. so you formed an ellipse ...

Now think of a roof, the 45* angle is your Hip run. Lets make the pipe bigger ... 12' tall Cut it at 90* you have a major axis of 12' & a Minor axis of 12' a big circle . 1/4 that pipe and you have the shape of your curve rafter ( exagerated for discussion) Now to get a matching hip rafter take the 12' pipe and cut it at 45* your minor axis ( diameter of the original circle has not changed) What you did change was the Major axis - it is now 17' ( the run a hip rafter has to make for a 12' common run) If you quarter the pipe , you will have a hip that will mate up with your jacks ... to make the jacks - you cut the sections of pipe to meet the length of the chord required for the diagional run of the jack - This is all easy to do on the construction master pro using the Circle & radius keys ..

Thats the concept... Now apply it to wood rafters... Your project will be a bit more complex since the chord of the circle will begin to rotate upward from your common fascia plane - that aspect will be dependent upon how much rise you add to the end of the hip.... Whew ..... but plotted - on paper all of it can be calculated

The drawing of the ellipse is also relatively simple .. But I'll leave that for another discussion ..

Are there any Asian roof framers that can jump in here ??
I'm sure there is more to it than we have discussed .

Hope that Helps Joe .. If you have any questions feel free to email me - I can fax some hand scratched drawings ..

Mike S

01-30-2005, 04:39 PM

What are you considering as a "simple curved roof"?

If, following your reasoning of the circle and the chord, consider this.

The length of a hip rafter is greater than a common rafter and a common rafter is longer than the jack rafters.If the length of the chords are the lengths of each rafter, hips, commons and jacks, and if you draw a line perpendicular and central on the chord to the out side edge of the arc, that distance will alter with each chord length.

If you then appy your method to that of pitching the roof in the normal manner then the top edges of the commons, jacks and hips will not come into a common plane other than at the ridge which is a fixed defined point.

Here is an exercise for you. Draw a circle, then draw a chord, yz
to represent the length of the hip rafterThat will create an arc yz. Then draw two more chords, parrallel to but shorter than the first one.ab and cd to represent a jack and a common.The sweep of the arc is constant, then visualise what happenes if you place the segments with point a and point c at the fixed point y with the chords ab and cd on the chord yz.

Let us know what you get.


01-30-2005, 05:33 PM

Another simple exercise with circles arcs and chords. Draw 3 circles, one with the diameter of the length of the hip rafter, one with the diameter of the length of a common rafter and one the diameter of a jack rafter.

Let us know what happens to the path of the curves.


01-30-2005, 06:15 PM
Chippy -
I don't mean to disrespect you in any way .
I have framed these flaired roofs ... Unless you have framed one, this discussion will be difficult to follow ..
The Commons , the Jacks are Chords of the SAME radius circle. you are just using shorter pieces of the circle .. Only the Hip is an ellipse . It is Elliptical because the Axis has changed .
Think of a regular roof , Nothing changes with the jacks & commons except the Length - The pitch - plumb & seat cut remain the same. All you are doing is selecting a a shorter section of the common rafter as it intersects the hip at a shorter distance. The Hip changes because it needs to travel a longer distance for the same height. The hip pitch is different , we merely adjust the square for the distance it has travelled on a 45.. It really has dropped it's pitch in relation to a common.
So the plumb & seat cut change , as does it's pitch.

SO if we use our radius to cut our flaired roof ... lets say 12' as I discussed earlier .. as we approach our jacks, they must remain on the same plane ... only they get progressively shorter .. The calculation is derived from the Chord length.. This is the strait line measurement that is equivelent to a common jack ... as the chord gets shorter , the radius never changes .. Only the circumference of the arc segment ( arc length ) changes ... The ONLY way to generate an ellipse would be to change one of the axis ... in the case of a hip, it skews from the common run .. therefore we get an ellipse. we get off our formula of 1-1 relation ship of minor to major axis .... unless you change either the rise or the run , the radius will remain constant ... Most of the roofs that use this , Have the radius on the bottom quarter of the roof, this is a common newengland porch detail & found on many of the local Dutch colonials , & english a few english tudor eave details

However - if the common rafter were laid out as an ellipse to begin with, than all commons & jacks share the same ellipse , & the hip becomes a ellipse with a longer major axis ...

Now in dealing with Joes proposed roof, this is more complicated - I'm not sure that the radius actually remains the sameas it approached the flaired hip - it could , but I'm not really sure how they actually plotted these roofs since I have never done one.

Joe ---
Gee aren't you glad we started this one..
Jump in here and distract us .....

Keep replys nice , I can edit your Posts .... :)

Mike S

ps .to all - be demonstrating this at JLC LIVE in 2005 workshops, so If you can attend , let me know if you will be there ...

01-30-2005, 06:55 PM

This has nothing to do with disrespect and non is taken, at least not by myself. If we cannot discuss matters openly and honestly which may offer a different perspective or correct any perceived or factual errors or offer advice then I see no point of having this forum. I am more than happy to continue this topic because it raises issues that are not necessary apparent initially to a roof framer looking at a roof.

What is your definition of a "flaired roof" a roof which has straight and in plane members from the ridge to the wall plate and the rafter tails only have a curved up swing. Or is it a roof that the hips, commons and jacks are curved over the total length of the top edge from ridge to facia. Different types of roofs different types of problems. A case of the difference between apples and oranges.

Have you tried the two exercises that I suugested?


Gary Katz
01-30-2005, 07:24 PM
I really don't understand what a mansard roof has to do with any of this. Maybe someone can draw a picture of the relationship.
Joe, this probably isn't what you're after, but while visiting the Gamble house a month or two ago, Brent Hull pointed out to me how the rafters were all different sizes. The barge rafters are probably 4 x 10, then the next ones are 4x8, then 4x6, then the rest of the commons are 4x4 or something like that. Obviously they borrowed the Asian look, too, from the Japanese Pavillion at the Worlds Columbian Exhibition. I think what you're after is much more pronounced. I'll hang around to see if one of the math heads (Joe Fusco???) comes up with a formula. I think Mike S. is right, though, it's something like a groin vault, huh?
Duh...thanks for 'editing.' I forgot the photo. It might be too small because of the size limits on the forum. If so, go to this page:

01-30-2005, 07:47 PM

Mansard roof, ask Paul Conrad post #11 or read my reply post # 12.

Can you explain where member sizes come into this topic, surely it is about shape and how to achieve that shape. I don't know whether you read my last posts or not but I would be interested in your comments.

The formula's and set out would depend on what type of roof you are really trying to construct. I guess we shall have to wait for conformation of roof type from Joe Wood and Mike Sloggatt

Where are Joe Wood and Paul Conrad now, they seem have gone AWOL.


01-30-2005, 09:19 PM
Chippy -
As I understand your excersice - It doesn't wash .. I tried to figure out what you were describing ... I either don't understand it or plotted it incorrectly oryou are incorrect.
Key question
Have you ever framed a flaired roof ?
I have - If you have you would know that the jacks are merely smaller chords of the same circle. I can - if i choose to waste materiel - cut 50 full common radius rafters, cut seat & ridge cuts. line tham all up on the roof. ( an excersise I learned from a European roof cutter) Than to establish my jacks , start cutting off the tops as they intersect my elliptical hip measuring from plate to hip - essentially adjusting chord length ... Period. No ellipses .. just shorter chords....

No matter what section of a chord of a single circle you wish to describe , whether it be 1" or 500" the radius nor Diameter does not change. PERIOD. A chord is merely a straitline measurement across a section of the circle. The things that change the chord are Arc length, & degrees of arc. NOT the Radius. Never... It's basic .
If the radius NEVER changes the layout method for the top of the rafter cannot change either. You will 'truncate' the length based on its chord length . THe only confusion I can see here is if we are describing different roof structures. Other wise I have built these roofs wrong for 28 years & never got caught .....


Tim Uhler
01-30-2005, 09:29 PM

Excellent posts. I'm going to copy them and save them. We have been framing some barrel ceilings and were going to do exactly what you suggested regarding the jacks. Once the common rafter with the curve is established, then I can use it as a pattern.

Don't know if you follow, but what I'm saying is the same as shown in the picture of the barrel hallway in Will Holladay's book.

Also, keep asking Chippy whether or not he has framed one of these. He has always dodged questions like that. I recently framed 2 barrel hallways that curved on the end. I layed out the "hips" like Joe Fusco said and basically like you said, and they worked perfectly.

Again, really good posts.

01-30-2005, 09:52 PM
Tim -
A barrel vault and a curved hip --- same thing only upside down. The plotting of the hip/ valley is the real skill. Unless you have framed one of these it is difficult to visualize much less understand . I don't know if you have a CM pro If you do - fool around with Chord length . the relationship of chord length to rafter length .. Strikingly similar ... the cool thing is that you can calculate the arc length & know the size of plywood or sheetrock... Something I have been trying to teach Myron Furgeson for some time now ,,,,,

For example :

I went into a Pizza Place in Virginia, It was called Authentic NY Pizza. Being from NY I had to try it, So I said to the Owner , HEY WHERE YA FROM IN NY ? He hesitatingly responded , Brooklyn ,, ( with a southern accent) I said oh yeah Where ? He responded 44th ST , by Yankee Stadium.... I said Ohhhhh really ? ....... I proceeded to Order a slice of Shenendoah County Pizza .......... Yankee Stadium is in the Bronx ,,,, obviously, he had never been there ....

I left , letting him think he had fooled me ...

Tim Uhler
01-30-2005, 10:02 PM

I have the CM Trig Plus III and I do use it for arcs. I like to use those functions to do arched garage openings.

Here is a pic of that ceiling. Jasen was using the 4' level to check and see that the curves fit. Fit perfectly. The hip was layed out as an ellipse.


Gary Katz
01-30-2005, 10:40 PM
I read through your posts, very carefully. I'm sorry but I didn't learn anything. I'm not sure why. Maybe I'm not bright enough. :>} Every book on architecture I've ever read describes a Dutch Gable exactly as Joe used the term--a hip with a gable on top of it--as pictured below.

As for different size members...that's exactly what the end result of Joe's roof is going to require: a different size ellipse must be cut into each jack or added to each jack, just like a groin vault. The example of the Gamble house is subtle but exemplifies the same progression, and the style originates from the same source, as I'm sure with your obvious knowledge of architectural history, you already know.


01-31-2005, 07:00 AM
Nice job there -

Side point - The hip rafter you have there should technically be Backed . It needs to go from a 45* angle to almost flat at the top. I've done it by using 2 - 3/4" ply for the rafter and tracing the ellipse from a pattern on the front side & the back side - off setting the lower on the rear side by the depth of the 45* cut... Hard to explain easy to do - If you lay it out in plan view you can see the cut you need ... It fades from the 45 to zip at the top

I see another workshop here .....

Tim Uhler
01-31-2005, 08:24 AM

Thanks. I know it should be backed, and that is why I use 1/2". I can avoid the process. It's early, but I'll read your post more carefully, because at some point, I'll do one that is too large to use 1/2". Thanks for the tip.

I have George Collins book. One my first visits here to the forum a few years ago, you recommended that book to someone, so I went out and bought it. It's been a very good buy. Another book someone recommended to me is "Construction Geometry" by Brian Walmsley. I had to order it from a Canadian book store online. It's pretty good too, because they don't layout curves using the string method, they are all plotted to get different curves.

Thanks again.

01-31-2005, 08:54 AM

We build barrel vaults a little bit differently. The 1/2" plywood chord or web (in your photo) carrying your 2x4's on the flat are installed every 16 o.c. We use 3/4 CDX for this (across width...short dimension) and instead of 2X on the flat we screw 1x3 every 4-5" (long way) o.c. across them. It takes more time, but just one 2X on the flat can twist and distort an otherwise perfect substrate for the layers of drywall. Metal framing is not a problem in this case. On larger barrel vaults we have to install joist to carry the span and build a web with several pieces of 3/4 CDX for each nailing point.

The only drawback to this method is depending on the radius and the clearance above the barrel vault, the insulators can have a tough time getting in there...especially when the vault is framed tight to the roof.

With most radius framing we try to "isolate" the rock from the 2X framing with CDX or luane depending on the radius and application. This tends to reduce most cracking problems due to lumber shrinkage later on.


Joe Wood
01-31-2005, 10:42 AM
Sorry guys, I'm not getting email replies.

Think I'm on to something. I'm busy Sketching, I'll check back in later.

Tim Uhler
01-31-2005, 07:24 PM

We build barrel vaults a little bit differently. The 1/2" plywood chord or web (in your photo) carrying your 2x4's on the flat are installed every 16 o.c. We use 3/4 CDX for this (across width...short dimension) and instead of 2X on the flat we screw 1x3 every 4-5" (long way) o.c. across them. It takes more time, but just one 2X on the flat can twist and distort an otherwise perfect substrate for the layers of drywall. Metal framing is not a problem in this case. On larger barrel vaults we have to install joist to carry the span and build a web with several pieces of 3/4 CDX for each nailing point.

The only drawback to this method is depending on the radius and the clearance above the barrel vault, the insulators can have a tough time getting in there...especially when the vault is framed tight to the roof.

With most radius framing we try to "isolate" the rock from the 2X framing with CDX or luane depending on the radius and application. This tends to reduce most cracking problems due to lumber shrinkage later on.


That is exactly the way we were going to do it the next time. Thanks for the tip. I'm glad to know in advance it'll work :-)

Joe Wood
02-22-2005, 11:08 AM
Sorry for the long delay, but I've been busy Sketchen.

Your talk got me to thinking, instead of trying to figure everything out beforehand ..

Joe Wood
02-22-2005, 11:09 AM
what I did was to first draw my roof planes (shown transparent here), with the exact curves I wanted.

Joe Wood
02-22-2005, 11:12 AM
Then I extruded each rafter up thru the roof plane and trimmed it to fit the curves at its’ top part, where it interested with the roof. Then I finished up the bottom of each rafter, and gave them all the same end detail. This roof has a 3’6” overhang.

Joe Wood
02-22-2005, 11:15 AM
Pretty slick. I don’t think I’ve ever seen an Asian style roof framed Western style like this, with rafters and trusses. I guess now that I’ve drawn it, I can go have full-sized templates made for each rafter.

Joe Wood
02-22-2005, 11:17 AM
I sure appreciate all the help I got for figuring this one out. Now all I have to do is find someone who wants one !

Joe Wood
02-22-2005, 11:37 AM
Shoot, I just posted 4 sets of images, and only one has shown up. Strange, I saw them come up as I posted them. I'll try again ..

what I did was to first draw my roof planes (shown transparent here), with the exact curves I wanted.

Joe Wood
02-22-2005, 11:38 AM
Then I extruded each rafter up thru the roof plane and trimmed it to fit the curves at its’ top part, where it interested with the roof. Then I finished up the bottom of each rafter, and gave them all the same end detail. This roof has a 3’6” overhang.

Joe Wood
02-22-2005, 11:40 AM
Pretty slick. I don’t think I’ve ever seen an Asian style roof framed Western style like this, with rafters and trusses. I guess now that I’ve drawn it, I can go have full-sized templates made for each rafter.

I hope I'm not double posting these over again. I can see my last 4 posts in the Topic Review section, but not in the actual thread. Oh well ..

Joe Wood
02-22-2005, 11:41 AM
I sure appreciate all the help I got for figuring this one out. Now all I have to do is find someone who wants one !

Joe Wood
02-22-2005, 11:46 AM
OK, just tried again and only one more set of images appeared. Strange how the two posts I made today are up in the thread and not at the end like they should be. Also strange is that I see them in the Topic Review, but not in the actual thread.

Can anyone else see the 9 images I just posted ?

I'll try again later with the other images.

02-22-2005, 04:19 PM
JOe -
Would you mind sharing the .skp file ? --- I'm sitting there trying to rotate your jpg.....
I'd love to see it up close


Joe Wood
02-22-2005, 05:48 PM
Mike! I've put too much into this to just send it out. besides, I'll probably be making and selling Plans of this, but for you, I'll strip it down some and then send it to you privatley since you're a fellow SketchUpper !

Lets see if I can post any more pics, then you can see some closer up.

Joe Wood
02-22-2005, 05:51 PM
Well, that worked ..

So, what I did was extrude each rafter up thru the roof plane and trimmed it to fit the curves at its top part, where it interected with the roof plane. Then I finished up the bottom of each rafter, and gave them all the same end detail. This roof has a 3’6” overhang.

Joe Wood
02-22-2005, 05:59 PM
Just tried to post more pics and they're not showing up. I see them in the Topic Review, but they don't show up in the thread. Back to what I was saying ..

Pretty slick. I don’t think I’ve ever seen an Asian style roof framed Western style like this, with rafters and trusses. I guess now that I’ve drawn it, I can go have full-sized templates made for each rafter.

I sure appreciate all the help I got for figuring this one out. Now all I have to do is find someone who wants one !

Joe Wood
02-22-2005, 08:29 PM
Sorry my replies here are so fragmented. I've tried posting these about 4 times now !

What I did was, after I had drawn the curved roof planes, I extruded each rafter up thru the roof plane and trimmed it to fit the curves at its top part, where it interested with the roof. Then I finished up the bottom of each rafter, and gave them all the same end detail. This roof has a 3’6” overhang.

Bob Dylan
02-23-2005, 07:42 PM

I'm a dinosaur using Cadkey 97. Wonder how quick I can do these frames with Sketchup.

04-01-2005, 06:26 AM
Gary Katz,

Sometime back I read a piece by an American architect concerning the use, misuse and incorrect terms used in American architecture and construction. He also made reference to the popularised misuse and propagation of incorrect terms even from well credentialed writers. The popular print and electronic media were also criticized for perpetuating incorrect terminology. Comment was also made concerning that much of the misuse of terminology was due to local terms and that there was no universal standardisation of terminology used in the US.

In your post you said “I read through your posts, very carefully. I’m sorry I didn’t learn anything. I’m not sure why. Maybe I’m not bright enough :>} Every book on architecture I’ve ever read describes a Dutch gable exactly as Joe used the term – a hip with a gable on top of it – as pictured below.” If, as you wrote you didn’t learn anything then you must already have knowledge of what I had written. Just in case you are unsure or didn’t understand what I had written I’ll put it in an American historical context, that way other readers of this thread can make their own minds up about the accuracy and origins of the terminology

Gambrel is a Norman English word, some times spelled gamerel, gamrel, gambril and gameral meaning a crooked or hooked stick. Gambrel is a stick or piece of timber used to spread open and hang a slaughtered animal by its hind legs. Gambrel is also the joint in the upper part of a horse’s hind leg, the hock.

Prior to permanent European settlement in America, Spanish, Portuguese, Dutch and English mariners and traders had visited or settled in to the area of south east Asia now the country of Indonesia. It was there they saw dwellings with a roof style where the end of a roof started as a hip and finished as a gable end at the ridge. The gable end was in fact an opening to allow smoke to dissipate from the cooking fires. This design of roof was brought back to Europe and the American Colonies and adapted to local conditions. The roof style is still in existence today in Indonesian rural communities.

The word gambrill was part of the Dutch language in 1601.

Various references are found in the original colonies in America about gambrel roofs including: 1737 Old Times, New England “One Tenement two stories upright with a gambering roof.”. 1765 Massachusetts Gazette “A large building with two upright stories and a Gambrel Roof.”. “Sometimes with the long sloping roof of Massachusetts oftener with the quaint gambrel of Rhode Island”. 1779 “The gambrel ruft house”. 1824 “In a Gambrel roof’d home”.1858 “a small farm with a modest gambrel roofed one story cottage”.

In the Dictionary of Americanisms by John Russel Bartlett ,1848, Gambrel, “A hipped roof of a house, so called from the resemblance to the hind leg of a horse which by farriers is termed the gambrel”.

The French or Mansard roof, dubiously attributed to the French Architect Francois Mansart, 1598 – 1666, but certainly a roof style used by him. In its basic form it consisted of a King post truss on top of a Queen Post Truss. This provided useable roof space as additional accommodation. The basic Mansard roof with gable ends was known a single Mansard roof with the roof having two different pitches, the lower (or pitch from the eave) being steeper than the upper pitch connecting with the ridge. A Mansard Roof which has hip ends is called a curb roof where the upper pole plates become a curb. The roof shape was varied and pitch proportions were modified over time to accommodate dormer windows and curved ends at the eave to reduce snow slip.

The French Mansard Style was again used when much of Paris was redeveloped during the reign of Napoleon III., 1852- 1870, this was also called Second Empire Style, in deference to Napoleon and the period in French history know as the Second Empire. During that period, whilst the roofs were basically the original concept of a King Post, Queen Post Truss they became far more elaborate in design and style. This building style was seen and copied by many overseas visitors to the 1855, 1867, 1878 and 1889 Paris Expositions which included Americans who were also exhibitors at them. The French Second Empire style of building was emulated in many parts of the world including America, both in public and domestic architecture.

The first recorded use of the misuse of the word gambrel to describe a Mansard Roof in America was in 1858. It was referred to as an American colloquialism for a Mansard roof. So, someone in America quite incorrectly started to call two entirely different styles and types of roof by the same name.

In your reference to the Gamble house you wrote “Obviously they borrowed the Asian look, too, from the Japanese Pavillion at the Worlds Columbian Exhibition(sic).”You are really drawing a long bow on that statement. The Gamble House looks nothing like the Ho-o-Den at the Columbian Exposition or the Ho-o-Do in Japan. For anyone wishing to make comparisons log on to www.gamblehouse.org and compare the photographs with the Ho-o-Den and Ho-o-Do photographs on www.geocities.com/SiliconValley/Orchard/8642/cs_hmgreene.html

A couple of weeks ago in a book store I came across a book called Dictionary of Architectural and Building Technology 4th edition by Henry J.Cowan and Peter R.Smith, published in the US by Spon Press 2004. Library of Congress ISBN 0-415-31234-5. That dictionary has a line drawing of a Gambrel roof which you call a Dutch Gable and a Mansard roof which you call a Gambrel roof.

For anyone who wishes to know what Dutch Gables really looks like and their origins log on to:


Your cad drawing looks good, however have you built it or produced drawings that any competent roof framer could use to frame it on site. If you have seen any Japanese construction or read any books on Japanese framing you would have been aware how complex their method of constructing or connecting joints really are. There is one big difference between American and European framing and Japanese framing and that is the size and lengths of timber being used. In Japan they use smaller lengths and sizes (smaller and fewer trees) and, have over many hundreds of years devised a unique method of joining short lengths together to form strong long lengths. In traditional building much of the framework is fully exposed for everyone to see. You cannot, as Henry P suggested “scab a piece of timber on top”, unless you are going to cover the framing.

Many of the methods of constructing and joints in framing are no longer taught or practiced. To make a current day comparison, when was the last time you saw a framer mortise a stud into the top and bottom plate? When did you see a framer even trench plates for studs? When did you last see a framer mortise and tenon a header into studs?When did you cut a tusk tenon into a floor joist or a bridle joint cut into a beam or rafter? Cutting and framing the roof you have produced by cad is not what your average framer would be able to do quickly and competently.


Joe Wood
04-01-2005, 07:38 AM
Geeze Chippy, you're getting too serious ! Maybe I should call it a irimoya roof, which is the style I'm really trying to get.

No, I didn't use traditional joinery in building it. Matter of fact I used western-style trusses, and created my own hybrid-style.


I'm hoping now that someone will want my buddhist-style curved roof, which really is a completly different roof style .. the oriental curved roof, framed western style, with rafters and trusses.

04-01-2005, 05:10 PM

The photos look good, but what I'm really intrigued about is the roof you showed a cad drawing of in post# 50 with the curved rafters.

I've built structures, modified South East Asian design, similar to the one in your photos for back yard shade including double pitches starting at the eave with a 10° pitch and then changing to a 60° pitch to the ridge using traditional western style stick construction without trusses.

By the way, that roof design in the photo which, as I said before originated in Java and Sumatra before the Dutch ever saw it is still being built there both in traditonal thatch and in corrugated iron roofs.

Serious, that is not a personality trait with me, quite the opposite. Correct information and terminology certainly with construction, well thats a totally diferent matter. If there is one thing worse than no information it is mis-information.


Tim Uhler
04-01-2005, 05:27 PM

Do you have any pictures of that work? I would love to see that and copy it someday :-)

04-01-2005, 07:36 PM

No, but if you are prepared wait a while I'll take some when I'm next in that area which 30km away. I don't think I'll able to do it before I go to the UK and Malaysia at the end of the month.
No doubt I'll take photos of traditional shaped roofs in Malaysia which are very similar to those in Java and Sumatra, that is unless they have domolished and changed the roofs styles since I was last there.


Tim Uhler
04-01-2005, 07:39 PM

Have a good trip. My brother and his wife visted the UK and Europe about a year and a half ago and really enjoyed it. I would love to see pics.

Chris Hall
04-15-2005, 10:55 AM
I am new to this forum, so greetings to all.

I have been reading this discussion with some interest as my carpentry specialization is Japanese architecture and roofwork in particular. Since that involves timberframe construction and joinery, there is probably not much point here discussing the details of how the curved hip can be framed, as this site is for people doing western platform framing with dimensional lumber, but I did want to address some of the information brought forth in the discussion all the same.
I think, Joe, considering what you are trying to accomplish using a western approach, you have come up with a workable design. Since your buildings are not Japanese in any way, other than by name, there is no point calling attention to the differences in what you designed to the real thing - and I'm sure you are well enough aware of these differences, which you termed "too complicated" - as the entire structural rationale in Japanese roof work is completely different than in Western. I imagine you will have no problem finding a client who wants that design, as there are few people in the west who would know the difference anyway, and to most, a building with a flared hip line is 'Asian'.
Regarding the picture of the Greene and Greene bungalow, and their attempt to mimic the upsweep in the plane of a gable roof, seen occasionally in Japanese gable roofs, using graduated thickness rafters, well, again, it's not how the Japanese carpenter would do it, and again, it is a decent solution given the medium. With widely-spaced rafters it is impossible to convey a really smooth sweep - it ends up a little clunky.
Chippy, who has provided some great clarifications on the origin and meaning of the term 'gambrel', also made the following comment which I'd like to address:

"There is one big difference between American and European framing and Japanese framing and that is the size and lengths of timber being used. In Japan they use smaller lengths and sizes (smaller and fewer trees) and, have over many hundreds of years devised a unique method of joining short lengths together to form strong long lengths"

Well, this is not quite correct. they do use smaller - and more numerous - scantlings in Japanese traditional framing, but it is not true that they use shorter lengths particularly, except perhaps in comparison to modern American timberframing practices. And they don't use smaller trees for most timber in the real work either, as boxed-heart timber is considered inferior. The only place you will see boxed-heart timber used is in the toko-bashira (alcove display post) and in o-bari, or the curvy spanning logs commonly seen in Minka construction. And in inferior-grade work. When boxed heart timber is used, a kerf is run along the spine of the cant to the center so as to allow shrinkage to be taken up on the one face. Quarter sawn material is greatly preferred.
When it comes to elaborate methods of scarfing timber together, the hands-down world champs must be the English carpenters who built the great churches in that country. Take a look at Cecil Hewett's book "English Historic Carpentry" to see for yourself. Japanese joiney in general is more refined and complex that western, but not so in the case of scarf joints (termed "tsugi-te" in Japanese). The French have an identical scarf joint to the Japanese in their traditional repertoire. I should also point out, as many of you are quite aware I'm sure, that no scarfing joint can equal the strength of a continuous timber, so it really is a technique of last resort.
I was wondering Joe how you came to view Japanese carpentry layout as so complex, when you are able to handle an autocad program? The level of technological sophistication and learning required to do or use cad programming, nevermind to actually come up with the various algorithms involved in the software, surely exceeds by a huge margin the degree of complexity in laying out a hip curve in Japanese carpentry, which, in its basic method, isn't so complicated at all. Really.
To grasp Japanese layout, it does help to be able to read Japanese, and that, if nothing else, is what they came up with the term 'fiendishly complex' for :^)

Chris Hall
04-15-2005, 07:19 PM
And one more thing I forgot to say in response to some of the earlier posts on this topic: the section of curve in a Japanese roof, at least (I can't speak for Chinese roofs), is NOT a section of an ellipse. There are some four usual ways to establish this curve and none of them involve producing a quarter ellipse (which would have way too much upsweep at the tip of the hip/kaya-oi intersection for Japanese tastes). Also, the curvature in the hip is not established on the hip and then transferred to surrounding jack rafters - rather the curve is established on the perimeter fascia (kaya-oi) and then transferred to the hip and jacks, as the case may be.

04-15-2005, 11:34 PM
Welcome Chris

Thanks for your Insight

I'd love to learn more about the layout techniques you reference ..

Any good books,

how about Photo's of yor Work ?


04-16-2005, 04:29 AM

I have read your two post with great interest and have now taken copies and added them to my collection of construction information.

Your knowledge of Japanese construction is far superior to mine and I accept your comments on my post. I should have been more specific in my comment and making of comparison between Japanese and European and American construction and material. The Native grown trees of the respective areas are different and it created variations of construction techniques. The size, length and volume of timber grown in North America is totally different to that of Japan and Western Europe.

The skill of the European and English carpenters and masons never fail to amaze me using such basic and primitive tools compared with today. The building that holds me awe has to be Westminster Hall in London built by William Rufus in 1099 and the "roof reno" undertaken by Richard II, 300 years later and is still standing today.

Anyone who has seen buildings built in their native environment, with native materials and by local methods has the ability to more readily recognise the corruption in design, naming and contruction in places other than their native locations. I count myself fortunate to have travelled and worked in number of countries including, Japan, South East Asia, South Africa, Europe and England so I am familiar and have first hand knowledge with the designs and types of buildings I have mentioned on this thread.

The Gamble house has to be a classic example of misdescription. Taking nothing away from the tradesmen who built it, it could at best be called an eclectic design and at worst a corrupted design. There have been many descriptive tags put on it.but non appear to be accurtate. It is certainly nothing like Japanese architecture and it is certainly not a bungalow and anyone who calls it either Japanese or a Bungalow has a very limited knowledge of those architectural styles. A bungalow, being a single storey lightly built house with a verandah. The design being adopted by the British in Colonial India. The word Bungalow is derived from the Hindi "Bangla" meaning "belonging to Bengal".

The ability of designing using autocad is only one aspect of construction, the other aspect is the ability to convert the design to practical application and construct that design. Joe come up with a cad roof design with curved rafters and hips, yet the structure he built was in fact a traditional western style stick roof.

There is a historical parallel with cad designers and skilled artisans who did nothing else but set out and calculate the necessary bevels, angles, curves for others to follow, copy and undertake the actual construction. There were tradesmen who did nothing but make templates for others to use and copy. The practice of particularly skilled tradesmen in joinery shops doing such work was still being practiced in my memory. The English term for doing that was making a "rod" derived from the time when a length of timber (a rod) were used to mark and measure repetitive measurements and before the time when all tradesmen had received formal education or could read and write). Templates were added to rods, but the name continued. I have had the pleasure to have known some old time served carpenter's and joiners (long since passed away) who worked in such an environment. In big joinery shops there was a place called a "rod room" where all the rods and templates were stored and issued out to the trades men needing them. A job sheet was given out to the joiner with a list of required timber and the reference to the templates or rods he would use to do the work.

Your information on the curves of Japanese roofs puts an interesting complexion on the topic, thanks


Joe Wood
04-16-2005, 09:43 AM
Chris, I also would very much like to see pics, and clear drawings of the hip details !

Chippy, I first designed (then built) a straight roof azumaya (notice I didn't call it a Dutch hip). Then I designed the curved roof version, which I'm still waiting to find someone to build it for.

I've actually designed about 5 curved roof versions now. Want to see something wild ? This roof has an almost 5' overhang ! I seriuosly doubt it'll ever be built (at least not in areas with snow loads and moderatly high winds), but it sure was fun to 'build' in CAD.

Joe Wood
04-16-2005, 09:45 AM
Two more images.

Joe Wood
04-16-2005, 10:37 AM
I think I have the long side figured out with those 3 trusses that cantilever out along the rafters (I could always add a few more trusses on the long side), but it's the short side (with those 2 2x4s that cantilever out) I'm not at all sure of.

Joe Wood
04-16-2005, 10:44 AM
More images of the trusses.

Does anyone know a structual Engineer in the San Diego area that I could take these drawings to, to have them engineered ?

It is a structual Engineer I'd want, isn't it ? Sorry, I've only needed to have my projects engineered a couple of times, and those times were years ago so I don't know any Engineers now.

Chris Hall
04-16-2005, 11:33 PM
Thanks Chippy for the concurrence of opinion on the Gamble house - I was somewhat apprehensive about taking on much-admired 'sacred cow' there, so it was nice to see that I am not the only one with issues about the masterworks. They Greene and Greene classic bungalows (the later James house, however remains my personal fave), have some wonderful qualities, but I really don't 'get' the exposed rafter tail thing or the cheesy torch-on roof. Inside there is that very sinuous timber work, though there is so much wood in most of the rooms of the Gamable and Blacker houses that they are rather dark. And then there is the furniture, some of which has become entensively copied now. The cloud lift is but one Asian motif, and it is a little odd how they chose to borrow that one and that it has now become so synonymous with 'Asian' - often done upside-down even. And the Greenes did go a little overboard with the fake pegging thing....
Anyway, that's another topic for another forum. As far as Japanes roof layout is concerned, I could give you several titles that deal specifically with Japanese layout, but all are in Japanese, and to order them you would have to be able to type in Japanese on a Japanese booksellers website. If you're up for that. let me know. Hida Tools in Berkeley may have a copy of one or two of them in the store. The only work in English which deals at all with 'kiku-jutsu', the system of exploiting relationships within the triangle and transferring them to the timber with the aid of the sashigane (framing square) is the book "Japanese Woodworking", which is in itself an amalgam of two earlier separate publications. The part of that book dealing with roof work is taken from one of the books written by Yasuo Nakahara on the subject. However, there are numerous problems with that translation, among them typos in giving certain key numbers, for example, ratios of scarf length to timber thickness (kinda critical), a chapter missing altogether (on Japanese treatments of the Western truss), and a section on aspects of kikujutsu which is so truncated as to be unintelligible. Maybe not quite as bad as the typical 19th century stairbuilding book, but close :^)
I am currently travelling around the East Coast of the US, so I don't have ready ability to explain much by way of drawings here. I have taught a workshop for the timberframers guild on Japanese compound joinery - last year - and I am looking to do some more workshops, but at present I have not sorted out much, as other parts of my life are in, shall we say, a state of flux. If one gets organized, I'd say please come and check it out if you want to get an intro.
Besides, I can't easily go about explaining how to lay out a curved hip without first explaining the way the Japanese go about the regular hip, and thet means explaining how they go about understanding triangles and deriving the pertinent cut angles from that, and that means also going into the basics of structure in a Japanese roof, the orientation of parts, their names, and so forth. It really is a huge topic. I have contracted with Linden publications to write a book on this subject, but that process is really in a nascent stage, so publication is at least a couple of years away at this point. Sorry.
As I said I am well away from my 'desk' at the moment, so I can't send you any pictures of roof work, either mine or other carpenters, but I can provide a link to a page depicting the assembly of a piece of furniture - a doctor's office desk I did last year (made of Honduran Mahogany and Maple) - which does, in fact, have 2 curved, very steeply sloped, hips on it. You will note that the hips, as well as all the parts used in the dogleg bend of the desk, are stretched by some 11% in width so as to allow for seamless unequal miters (itself a internally wedged joint) at the connections on the upper table frame rails. The desk is 9 feet on one side and about 6 on the other, is assembled completely with joinery - no glue or metal fasteners - and comprises some 260 parts and took me a ridiculous, and 'unprofitable' 1100 hours to make, and three days to assemble. The pictures are unfortunately too large to fit your screen, so you will have to scroll your way around, and there are some 115 pictures, so if you have a slow collection, well, my apologies. Take a look (I'm the guy with the ponytail): http://joinery.6x.to
One point I wanted to add to what I wrote yesterday is that the Chinese and Korean curved roof is much more pronounced in curvature that the Japanese, and that their method of accomodating such huge upsweep at the hip is to actually tenon on a vertical member towards the bottom of the hip that catches the perimiter fascia some distance above the hip itself. This is not done in Japanese work. Also, the fundamental difference between Japanese roofs and their Korean/Chinese antecedents is in the Japanese development on the double roof, which went on to allow a number of innovations in the rest of the structure.
The best reference book in English on the development of the Japanese roof is the out-of-print book by architect Mary N. Parent, "The Roof in Japanese Buddhist Architecture". An excellent overview, but no technical specifics or layout info.

04-17-2005, 08:05 AM
Wow Chris -- Nice work .
Now I can't wait to see the roof frame pics.

You should write a primer article for JLC on The concepts of Japanese roof framing - Or even post an article on the web - I would love to understand more of this.

The desk is beautiful, to bad it has to have phones & wires on it .


Maybe we should start a new Thread on this ? -

Joe Bartok
04-18-2005, 10:37 AM
Chris: Very impressive, and more than a little overwhelming!
That's not just woodworking, that's art. How long did it take you to master the intricacies of this kind of joinery?

Chris Hall
04-18-2005, 08:55 PM
Joe - thanks for the compliment, however I must deny that what I do is 'art', along with any claim whatsoever that I have mastered any aspect of Japanese carpentry. It's more like a road filled with the potholes of errors than anything else, and i continually astound myself at new ways of making mistakes. I could show you numerous errors in that desk project. The maple grille bars for example, which I attempted to fan into the hip rafters, ended up looking goofy. I realized afterward, with that ever perfect hindsight, that the way to fan them would have been in two directions at once rather than simply leaning each bar to one side. I had thought of doing that initially, but the time constraints, and appreehnsions about the complexity of milling compound curves in the bars, 'forced' me to drop the idea. In the end though, I tend to think that another day or two of work is really nothing compared to the amount of time the finished piece sits there getting stared at. No one, after all, seems to remember a late delivery, only a bad one.
One reason I am drawn to Japanese carpentry is the depth of it - the more you learn, the more you realize there is to learn, like some sort of ever-receding horizon which you can never hope to reach but one that I feel nonetheless compelled to move towards. I'm never going to be as good as someone who apprenticed as a child with hand tools and has kept at developing their skills with dedication, and that is okay. So, the idea of 'mastery' of Japanese joinery is a concept that makes me laugh.
I lived in Japan for 5 years, but I am largely self-taught in Japanese carpentry and it is ideal to me to seek instruction wherever and with whomever you can find it and then just go at it with trial and error. Isn't that what being a journeyman is really about? Right now, for instance I am out on the East Coast so that I can take a 4 day workshop at the Heartwood School with French carpenter Boris Noel, in "Drawing in the Way of the Campagnon". The French and Germans use a different system of developed drawings (than the British and American systems) to figure out complex roof problems, and I anticipate learning a huge amount. Well, I hope not to be totally confused for most of it anyway.
Traditional French timber carpentry, which is primarily roof and stairwork, can be complex and even whimsical - - like helical turrets and onion domes with the purlins running in spirals, akin to handrails in a geometrical staircase if you can picture it - really mindblowing! I don't know why they do things that way, but I'm glad they have figured it out anyhow. I think the world needs a bit more freaky/imaginative stuff in carpentry, so long as it is well made and structurally sound. Anyone care to discuss that idea?

04-23-2005, 07:31 AM

I like your work and you have amply demonstrated that different nationalities and cultures have different approaches to similar products and it is not a simple task of copying but one of learning and understanding.

The Gamble house is a case of "Perpetuating the popular misperception". There are many more buildings and architects that deserve more recognition.

For a smart lateral thinking architect I believe that Sir Christopher Wren has to be in the major league. Before he built St.Paul's Cathedral after the great fire of London he built a public house in Fleet Street called the "The Bell" for his workmen. It was far enough away from the site to prevent workmen from leaving the job for a quick drink, but close enough for them to walk after work to quench their thirst. Hopefully I'll find time in the next few weeks to revisit "The Bell" and sink a few pints myself. The other smart move by Wren was to take a house on the opposite side of the River Thames so that he had an uninterupted line of sight of work in progress and could get there very quickly by a boat ride across the water, the fastest way to travel London in those days.

Joe Wood,

I guess you can call any structure you build by whatever name you choose, however sooner or later you will come across a person who has sufficient knowledge of building and architeture to recognise your error and thus compromising your reputation and integrity. There is no such thing as a Dutch Hip and the structure you built looks nothing like traditional Japanese architecture. It does, however look like many structures seen from the Indian Ocean, through the Indonesian Archipelago into the Pacific Ocean as far south as The Cook Islands. They all have different names, depending on the language, however the translated generic English term is pavillion.In parts of Indonesia it is called a Bale.


04-24-2005, 07:18 AM
"There is no such thing as a Dutch Hip"

A dutch Hip is a common term here in the States for a Hip roof -With a gable atop - It is used for Ventilation purposes - There are other variations , Boston Hip- snub hip etc. These are variations areused to solve the Cooling needs of the attic space of the Hip roof. THey are commonly known & recognized at least anongst the East Coast Framers. Which I might point out, East Coast & west Coast terminology & techniques can be vastly different. We have 'West' Coast saws & 'east' Coast saws .

Sort of like the first time I called a guy a Bum in front of My English friend. We were both referncing different concepts ...


04-24-2005, 09:35 AM
"There is no such thing as a Dutch Hip"

I guess it's another example of "perpetuating popular misperceptions". A roof starting as a hip at the eave line and ending with a ventilated gable at the ridge was being built in what is now known as the Indonesian Archipelago well before the Dutch settled there and well before European settlement in North America.(read post #55)

The original American colonists called such a roof a gambrel roof so the corruption of the name is "home grown". Changing a name to describe a known entity is a bit like trying to reinvent the wheel and claim originality. In lexicology it is stated that many words are corrupted and meanings changed due to lack of education and knowledge and it is certainly not the sole domain of the USA.

Obviously being able to differentiate between colloquial, slang and formal words and terminology would minimise the misuse of words.

I wonder whether a roof ridge capping should really be called a "Dutch Cap"


Chris Hall
04-24-2005, 10:01 AM
This discussion of terminology vis, "dutch hip", "gambrel" and so forth is interesting and informative. While taking the course in French drawing over the past four days, I asked Boris, the French compagnon teaching the course, what he called a double pitched roof, the lower pitch of which is more steeply sloped and is laid out commonly within a circle (in elevation) - he said, "a mansard". I then asked him what they called the same sort of roof with hipped ends, and he said, "that's also called a mansard". I told him that in N. America, the first roof was called a 'gambrel' and he said he had never heard the term before.
Although the Unabridged Random House Dictionary shows a picture of a double pitched roof for 'gambrel', I am inclined to agree with Chippy that there has been some mistake made along the line in the history and use of that word. Nothing new there, the English language is full of similar instances, like the Spanish 'norangia' (meaning, orange, the fruit) becoming 'a norange' and then by vowel migration, 'an orange'.
The problem is that is we change the name of a double pitched roof to something other than 'gambrel', in N. America at least, a lot of carpenters are not going to know what you are referring to....
By the by, the French call a bevel gage 'sauterelle', which means 'grasshopper' (presumably in reference to the hinged leg of the insect resembling the tool). I love things like that!

04-25-2005, 03:22 AM

Thank you for passing on the knowledge of your French compagnon, Boris. A Frenchman should know his own national architectural styles better than us "Anglo's".

Now I know the origin of the name "Grasshopper" however that name was given to a Handrail gauge used for gauging lines on handrail wreaths.Corruption by "Anglo's" again.


Chris Hall
04-25-2005, 08:48 AM
I wanted to make a correction to something I wrote earlier. I was staring at a book of mine that deals with the various methods used in Japanese construction to lay out a curved eave, moving along the eave edge towards the hip, trying to consider what shapes they laid out geometrically-speaking, and I realized that one of the methods does in fact develop a section of an ellipse. It does it in an unconventional way to any of the methods I would normally use to form an ellipse, so I hadn't seen it clearly as an ellipse section.
Looking at the other methods, one forms a parabola section, one forms a sine curve, and another a section of ellipse. Another yet is similar, except the normally vertical axis to derive points on the ellipse is inclined, and I'm not sure if that makes an ellipse or what.
The different methods used give slightly differently-shaped curves, so it could be fairly said that any number of methods could be used to develop any type of curve desired. Again, in Japanese work, where the curve reaches the hip it never approaches a vertical orientation, as normally seen in Chinese architecture.

Joe Bartok
04-25-2005, 10:42 AM
Chris: I've always loved working with all kinds of math, but have never had the opportunity to implement it in practical terms on this level.
Hopefully there will be more threads in future regarding layouts involving mathematical curves. This one's certainly been interesting.

04-25-2005, 02:16 PM

How did your sawhorses turn out?

I was doing some log truss joinery instead.

The greman/french course looked fun.


Chris Hall
04-25-2005, 04:49 PM
In reply to Lumpy,

the compound joinery course for the TF Guild went really well last year, judging by the enthusiasm of the students and my own expereinces struggling to convey the material in as clear and concise a manner as I could. One of the students, on his own initiative, did a 5 page write-up on his experience in the course, and that appeared in a recent issue of Scantlings. The only regret I have from the course was that two weeks wasn't enough to cover any of the material for the three projects in adequate detail. The sawhorse, which I allotted 3 days for, really needs a week. Ditto for the hip rafter project. Maybe next time....

The French drawing class was really good - thanks for asking - and I'm looking forward to working on some of the material soon. Descriptive geometry, after all, was invented by a Frenchman, so I knew there would be some interesting techniques in the French method, and I definitely wasn't disappointed in that regard. I thought Boris did an excellent job, particularly in light of the fact that he was teaching the material in English!

Joe Wood
04-25-2005, 05:21 PM
Chris, even though I figured out how to do a curved roof, Western style, I really want to know how to do it Asian style.

I've really tried hard to understand how they do it, but I'm still completly in the dark. I even got that book, The Roof in Japanese Buddhist Architecture, a few months ago, and it left me even more confused.

When you get back home, could you do some drawings, or post some pics for me that hopefully will help me understand ?

What I'd really like would be some CAD drawings that I could poke around in. My CAD program is SketchUp which you maybe don't know, but I can import dwg files into it.

Sure would appreciate it if you could help me understand.

Chris Hall
04-27-2005, 08:46 AM
Hi Joe,

well, as I wrote above, I feel it is going to be pretty tough to convey this sort of information in this format, particularly as I lack the sort of drafting software at the moment that would permit me to import drawings into this thread.
The book you got by Parent is, unfortunately, not going to be of any use as far as learning how to lay out a curved Japanese hip. Did you mention earlier in this thread that you had other books on Japanese layout? If so, please tell me which one(s) you have. If you can't read the title, just describe the cover to me and I will probably know which one, and we could likely look at a particular page at a time as I have most of the layout texts available in Japanese.
And if you have a Japanese layout book of any sort, and have been trying to work out the method - starting just with the straight hip, mind you - then I can ask you one question at least that is important to this problem, and is fundamental to Japanese triangle geometry: what is "chu-ko"? If you know the answer to that, then we might have a starting point from which I could try to explain some of the layout. If not, well, maybe there's some other way, but I imagine it will be tough sleddin'

Joe Wood
04-27-2005, 09:55 AM
Aw heck Chris, I was hoping you could post some simplified drawings. What I really need is to be able to just go and see an actual roof built with the curve, then I'd understand it. Don't suppose you know of any around the So Cal area I could go and look at ?

The only books I have about Japanese roof framing are, Measure and Construction of the Japanese House, and, The Japanese House, and neither goes in to curved roof framing.

Which books would you suggest I get ? Sorry, but I don't speak or read Japanese.

I have no idea what Chu-ko is. Looked it up with Mr Google but all I found were references about a Mongolian repeating crossbow, and a Chinese scholar/statesman .. nothing about roof framing.

Chris Hall
04-28-2005, 03:42 PM

reading your reply, I have to respond to your comment that seeing "an actual roof built with the curve and then [you] would understand it" - uh, I don't think so. You've already posted pictures of a Japanese temple roof, and that didn't seem to help you much, to be frank. An example in front of you is unlikely to give you much insight into the geometry of layout, and how curves are transferred from one plane to another.
Your books are, as you mentioned, pretty scant when it comes to roof framing. The Engel book, which is well detailed in regards to other aspects of the Japanese house, is strangely mute when it comes to the Japanese roof, which is the most significant aspect of the architecture. The word for roof, "ya-ne" after all means "the root of the house". I was half expecting you would have the book "Japanese Woodworking", which gives a little coverage to roof work, and is more confusing than anything else. The books on Japanese roof layout are all in Japanese, so you are out of luck unless you are willing to study the language to some extent. I am writing a book on the topic, but it will be a couple of years yet until it is ready. Sorry.
I find it a little curious that you are trying to emulate an architectural form you apparently have so little first hand, or even on-paper knowledge about - what's the deal with that? If you aren't interested in doing the work in the honest manner in which it is executed, ie, with real joinery and correctly proportioned timber elements, then why bother doing it at all? Working at it from a 2x stud lumber, platform-framing perspective, you are never going to get remotely close to recreating the real thing, any more so than trying to emulate concrete or steel with wood.
If you are interested in the real thing, and not just a distant copy, then I am happy to help, but it is pretty complicated, and would be a real challenge to convey online, even with pictures. If I posted a descriptive geometrical drawing of Japanese curved hip roof layout I can assure you it would not help you much, regardless of how much autocad you are used to looking at.
As far as buildings in California go, there is a garden pavilion in a Japanese garden in both San Francisco and San Jose, but I don't recall either having curved hips. There is the Ellison residence in South San Francisco, and if you can get past the 300+ cameras, and armed security, then you could get a glimpse of many curved hips on the buildings there.

I wish you all the best.

04-28-2005, 05:34 PM
Chris -
The Asian Architecture has not had as great of an influence on us here in the US as European. I think Joe has taken some of the elements he has seen and copied their concepts in making a western frame version the concept. It's obviously the elements that have inspired his homegrown design. Really whats wrong with that if he has clientele that like it. The moderation of the design might make it more appealing to the western consumer.
To the trained Eye , well thats another story. I suppose if a Japanese business man hired him to build a pool house in the backyard, he might be in trouble.
I admire Joes eagerness to learn and understand , to perhaps perfect his techniques. I too have a desire to at least understand the Japanese methods ,as it may very well simplify some aspects of my work.

Most of todays Architecture is a mixed bag of Influence. Who sets the standard for what is technically correct ?

The bottom line is the client. If a client sees Joe's Asian influnced western framed gazebos and is willing to spend 20,000.00 to have him build it. THan it is excatly what the client wanted.

If you want to be a purist, Thats ok , . But how far do we take it ? Do we refuse to Put electricity in a home designed after 18th century architecture ?

It's guys like Joe that fuel the designs of the future. I look forward to seeing what he comes up with as he learns more about it ...
I look forward to your posts as you shed more insight on this subject.


Joe Wood
04-28-2005, 06:09 PM
OK, if that's the way you feel about it. Myself, I'm just trying to better understand this roof structure, so that I can come up with a better Design, maybe Style of my own, not a Copy of something else. I mean no disrespect, to you or the Traditional Art, but I'm not trying to copy this style, I'm just trying to better understand it so that maybe I could simplify it, or maybe even take it somewhere else.

I do know quite a bit of joinery, but you see, in the style I want to build, I don't want to depend on alot of joinery.

What am I trying to do anyway, make a hybrid out of Eastern and Western styles ? Guess so, and you're wrong about actually seeing one making me understand it better .. well, I'd like to see, and crawl around in, a couple of them, then maybe I'd finally understand it. Or give me 5 minutes, with an accurate CAD model of one, and I betcha I'd finally understand. I just need to See how all the parts are joined together.

Maybe I won't even use any of these elements I suddenly understand, I don't know. I understand how to do the irimoya style roof, with all the proper joinery blah blah but there's something I still obviously don't understand about the same style but with curved hipps and draped eaves. I think I'm almost there though ! Gettin closer anyway !

Chris Hall
04-29-2005, 02:58 PM
Mike - thanks for the comments, and sure, if someone wants to pay Joe x-dollar to build them something, and they like it, then of course there is nothing in the world wrong with that.
I don't know what you mean exactly by 'technically correct' as it applies to Japanese traditional wooden architecture, as it is an evolving medium of the carpenter's expression, but I do know that it is an architecture based on carefully worked out (yet varied) proportioning systems using timber components, not 2x stock, and rarely if ever in buildings with the wooden structure open to the weather.
As far as being a 'purist' goes, I don't know if you were suggesting that i am one, but if so, I can assure you that it is not the case. There are many 'defects', fairly widely recognized, in Japanese buildings that I am not intererested in replicating for the sake of 'authenticity'. I do wish to preserve the tradition and approch to working with wood that the Japanese have developed over millenia. Of course, it is a type of architecture that is very expensive to duplicate, and at odds with modern engineering ideas, building codes, and so forth, so compromises are inevitable most of the time.

Joe - I appreciate that you are not trying to copy Japanese carpentry, but I wonder why you use Japanese names for your buildings then? It's like going in an "Indian" restaurant in Japan (many of them) and ordering 'curry' - it is a Japanese idea of what curry is, for Japanese tastes. I personally think they should call it something else. Of course many Japanese think they are eating authentic curry. I hope my analogy is clear. I'm certainly not going to tell you what to build or how to build anything, but if you're going to use Japaense terms for what you build then i feel obliged to point out that I don't find it Japanese at all.
If, as you claim, you understand the irimoya roof, then I will call you on it - what is the typical range of slope of the bargeboard? Measured along the theoretical hip line, from ridge to eave line intersection, what are the common proportions of size of the gables in relation to that length? Where do the hip rafters terminate in the roof? Are the hips usually 45-deg. or irregular? Can you have an irregular hip that appears 45-deg. under the eave?
By the way, whay would you need to see how the 'parts are joined together' in a curved hip if you aren't using joinery? I think that the issue is not how the parts go together, but how the shapes of the parts are derived, and how those shapes are transferred from one part to another.
When I was asking you before if you knew what 'chu-ko' was, I was wanting to know if you were familiar with it, not whether you could do a google search on the term, as you would need to know that both vowels are long sounds and that affects the spelling. There is probably very little on line on the topic anyway. But if you knew what chu-ko was about, we would have a starting point for discussion at least - without it, well, I would have to get even more long-winded than usual, and, I just don't have the time or technical resources to make that possible online.
Well, it seems like a moot conversation at this point. I am sure you will figure out a solution that works for you, and will delight some client I am sure. Best of luck.

Richard Birch
04-29-2005, 06:24 PM
Sounds like “Crème de BS?” to me!

How do you say “Where’s the meat ?” in French?

“Moot”? I thought you said ”le Moose.”

Anyway, lets get it on!

What is the concept behind Japanese Roof Architecture? Let’s get to the root (f) of this ……Japanese …………….?????????????

Please don’t vote me off the show/Island, Hee=Haw!!! You know what I mean?

Chris Hall
04-29-2005, 07:55 PM

your input to this discussion is...different, I'll give you that. It is difficult in this medium to know whether someone is trying to be humorous in a light-hearted way or something else.
I'm not sure what to say to most of what you wrote, since it is pretty much nonspecific, but I think I can say something to you question of, "what is the concept behind Japanese roof architecture?": well, the purpose is to cover a building's structure and protect it from the weather, to convey certain cultural ideas about beauty and function - in short, the 'function' of the roof is pretty much the same as in any other part of the world.
The unique invention of the Japanese roof, as compared to its Chinese and Korean antecedents, is the invention of the double-roof, in which there is a lower roof visible under the eaves that is pretty much decorative and a hidden structural roof above it that carries the weight of the roof and resists loads imposed upon the roof. This invention allowed the free positioning of interior structural elements, and made the space under the eaves a lot more pleasant, plus decreased rot and mildew since the decorative roof is at a shallow slope - usu. 3.5 in 10 - allowing for better air circulation.
Is that 'meaty' enough for you? In reality, the above is only an appetizer - just generalities, in response to a general question. If you have some more specific questions or comments, please post them.

Richard Birch
04-29-2005, 09:47 PM

Different? Me? This forum has been lacking “Spice”. If your going to “Talk the Talk” you better ”Wok the Wok” I thought the “Wok” lacked ‘Talk”. Let’s fire this baby up. Pardon my “Americanese.” (Getting our world mixed up)

“Is that 'meaty' enough for you? In reality, the above is only an appetizer - just generalities, in response to a general question.”

Exactly! You give us “Vienna Sausage” and tease us with "Nikujaga". “This ain’t a fishing trip, we’re after meat and potatoes.”

See=Saw, if you know what I mean.

Now let’s layout a curved common and transfer it to the hip, Nipon style. Shape first, joinery second. (English prefered).

Chris Hall
04-30-2005, 01:37 PM
Well, Richard, I understand your impatience with not getting what you want on this forum, but I will mention, in an effort to clarify, that this thread originated with Joe Wood out of his desire to find some answers as to how to accomplish a simulation of an Asian curved hip/eave using N. American framing techiniques. He seems to have come to a solution that works for him using a CAD program. I am interested in the topic of curvature in roofs, and have laid out and constructed roofs with curved hips in the Japanese manner, and when i came across this website and forum on this topic, i was compelled to participate. This does not however make it my job to teach you how to layout a Japanese roof - that is beyond this medium, or at least my capacities to utilize it in that manner, and more properly belongs in a forum of Japanesetools.com or the TF Guild forum. I work as a timberframer and a furnituremaker, not a stickframer. It's like chalk and cheese trying to mix the two, so giving you all the details is largely a pointless exercise.
The above being said, I gather your intimation that I may be all talk and no walk, and in response to that I would agree with you. This forum, is, after all, just an avenue for talk. If you want to learn how to do this sort of curved roof layout for real, then you are going to find that you will remain frustrated by whatever I might be able to bring to this thread. I've spent the past 5 years deciphering Japanese layout from textbooks, made both models and buildings to cement my knowledge, made numerous mistakes along the way, and at this point realize I've only just begun learning, considering the depth of material. If you figure i can just throw it out on an easy-to-digest platter for you to feed on, you are dreaming.
Of course, I can imagine your impatience with the above, so i will explain then, pointless as it may be, how to do the layout. I doubt it will be of any help to you or anyone else reading this, but you asked, so here goes.
Firstly, your entire concept of how the layout is done in "Nipon (sic) style", vis, "let’s layout a curved common and transfer it to the hip", is incorrect. I mentioned this earlier in the forum - have you read the entire thread, or did you just jump in at the end? Since I don't have the means to transfer graphics to this thread, then I'll simply reiterate what I said earlier. The line of curvature is established on the kaya-oi, and then transferred to the hip rafter, and possibly the keta (wall plate) if the curve originates before the intersection of hip and keta at the corner. If the jacks are to be curved, then their curvature also is derived from the curve of the kaya-oi.
First, therefore, you need to determine the origin of curve for the kaya-oi, the eave projection (measured from the centerline of wall to the lower front arris of the kaya-oi, and the rafter spacing. Then the amount that the kaya-oi rises is established. The roof is drawn in plan view and suplemental hinges are added to give a view of the side of the kaya-oi and the side of the hip rafter.
Since the edge of the eave rises along the sine of the common rafter triangle, and the curve on the kaya-oi is laid out, by various methods, as if the kaya-oi were in plane with the groundline (but in the roof the kaya-oi is in plane with the roof surface like a typical purlin in Western timberframing) the transfer is accomplished by taking the chu-ko measure, at each rafter centerline, and transferring it to the hip in side elevation. The hip has a lines laid on it representing top arris, line of backing cut, which is the top surface of the jack rafter, and a line giving the bottom of the jack rafters. The centerlines of jack rafters and their mortises are then laid on the hip.
Proceding from the origin point of curvature on the hip, then, the chu-ko transfer measures are added to the line that gives top of jack rafter, at each rafter centerline, thus establishing the line of curvature for the hip. Of course, the line giving bottom of jack rafters must also be brought into parallel with the curve line.
So, that, is the overview of the process, and that is for the simplest of examples, the single hip, and no double roof. This sort of roof would only suffice for some tiny structure like a shed. Also to be given are the cut angles for the kaya-oi where they meet on top of the hip, along with angles to give the joint layout for that, and the angles to give jack rafter tenons (ie, the clip cut and top surface angle), and additional splines runnning through hip and jacks, especially the hanging jacks, and the nose cut treatment for the end of the hip rafter must be decided. Of course, overlaying the kaya-oi are other timbers, such as the urago and yodo, that need to be laid out to fit the curve of the kaya-oi, and meet on top of the hip above it.
Did that help you out Richard? I gave plenty of specific information, so run with it. Let me know how you do, and if you have any other specific questions that make you want to 'fire this baby up' then by all means please post them. I've walked my talk, so what of any use to this discussion do you have to contribute?

Richard Birch
04-30-2005, 05:13 PM

“I've walked my talk, so what of any use to this discussion do you have to contribute?”

Put Up or Shut Up, Eh?

Wee! Qui! Si, Senior!

Don’t get me wrong Chris. Young men with your level of interest in the kind of Carpentry you are involved in don’t come along every day. You’d be the first for here since I’ve been participating. I hope to see more work from you. You write so eloquently too. Got to keep you honest though!

I am not a Japanese roof framer, never will be, and may not be able to offer much first hand constructive contributions on this specific subject. I don’t speak, read, or think Japanese either so the concepts may not ever translate. But if something I say can inspire a person to do better, either at work or here on this forum, and even though it is said in a manner that upsets at times and is humorous at other times, it is always meant to excel the times. “Fire it up” You know………? I did get a little bit better explanation about the layout of the Curved Hip from you after all.

(As you noted) I think another problem with the roofs you (and Joe Wood) write of, and I read here, is that I have a preconceived concept of “Japanese style” curved roof structures to overcome. I see Japanese roofs as “Organic” in nature and truly free form 3-D. Designed by nature and understood by man. Akin to wooden ship building with it’s intersecting elliptical and irregular parabolic bowls. Like sections of different sizes bubbles or bowls arranged to provide a skinned rib cage for a roof shelter or a ship. If you could put the design idea in words that convey the “Concept” better, maybe that would be the grand panacea to understanding these designs. For example: I like to convey the concept of “Western” style home building as “A Glorified Box”. If my new employee has any background in plain geometry this simple “conceptual” statement seems to help them use what they already know. But if they don’t have any geometry then they go nowhere quick. (Which is where I might be going there with this subject myself, lacking any background in “Japanese” geometry. But I brought matches!)

After reading your last post it seems to me that scale model building would be big in the Japanese style roof design process. Using paper models seems likely. (What? Another (?) Pre-misconception?)

As far as getting the design basics across to me and the other readers of this thread with illustrations why not just scan a hand drawing and post that? I’ve done it. Crude, but effective.

So, the ridge is curved, the fascias are curved, the hips are curved, the jacks have graduated curves, but there is a flat area in the middle? How do you lay that out simply?

Dreaming? Yep! Always.

Chris Hall
04-30-2005, 09:43 PM
I am now thinking that I am spending too much time here writing on this forum, but, well, here we go again.
Richard wrote: "So, the ridge is curved, the fascias are curved, the hips are curved, the jacks have graduated curves, but there is a flat area in the middle? How do you lay that out simply?"

The ridge is not curved in Japanese architecture. This is common to Chinese roofs and is a sort of 'defect' that comes about as a result of dealing with a curved bargeboard. The only "Japanese" buildings you will see a curved ridge on are those that are specifically Chinese in origin and are being preserved as original. Read "The Roof in Japanese Buddhist Architecture" for a fuller explanation. On gable roofs, sometimes the gable end is upturned, and this is accomplished by curving the supporting purlins. More commonly however, there is a 'minoko' (the 'scythe'), which is a roll-down at the gable end of the roof plane. This is also effected by curving the purlins.
The jacks may have graduated curves, but that is a fairly uncommon way to do it, especially in single layer eaves. Usually, and obviously more simply, the jacks stay straight and become progressively more parallelogram-shaped as they move along to the corner of the eave to the hip rafter. since both the hip and the kaya-oi rise together, the jacks can run straight. If the curve begins before the keta crossover, then the keta must also rise to meet the jack rafter.
As for a flat area in the middle, I presume you mean in the middle of the eave, halfway between hip rafters. If that is what you meant, then yes, that is typical in Japanese roofs. In Chinese and Korean roofs, the whole plane of the roof is curved - well - it actually looks 'sagged'. That being said, there are also Japanese roofs with the curve to each hip originating around the middle as well. The don't look saggy though.
Richard wrote: "After reading your last post it seems to me that scale model building would be big in the Japanese style roof design process. Using paper models seems likely. (What? Another (?) Pre-misconception?)"

Yeah, scale model building is often undertaken for temples, and in the case of some temple razed during the bombings of WW II, the models are all that remain of some structures. The models are extremely detailed and costly to make, often taking a year to build, typically in 1/10 scale. This may seem unbelievable, but in the case of temple projects that can run to 10-15 years ,it is not so surprising to invest in a detailed model. The models can cost as much as a house! Paper models are often used, though, as far as I know, most often by teahouse carpenters to experiment with room layout, and not roof layout.

As far as scanning a hand drawing and posting it, that is not so convenient to do in the place that I live, being a small island, and my current schedule doesn't permit the time, but maybe in a few weeks I will be able to. by that time, this thread may be dead, who knows?

Richard Birch
05-01-2005, 09:03 AM

I look forward to seeing a few hand sketches from you if you find the time and means in a few weeks. I do not think this “Lumbering Giant” of a thread will die by then. It may rest and take it easy for a while but it is far too interesting to just dry up completely.

“Fiendishly Complex” a notable comment that sticks out profoundly! One of the other things that visually sticks-out about these roofs shapes is the “apparent” volume of space created by these “Fiendishly Complex” roofs. It’s as if they expand space. I did a little perusing on the net this morning hoping to learn more and noticed that some of the examples of the architecture were relatively small but the impression I was overwhelmed with was illusion of “Space” the roof design creates. It is also noticeable in Joe Wood’s sketches. It was an interesting phenomenon that had not realized before. Is this something that is taught to students of this type of architecture? It would seem to imply that certain proportions must be maintained to maintain the illusion. Things like basic pitches and curves. (?)

Another thing you mentioned of interest is “the sashigane (framing square)”, and also the 1/10 scale models. I know nothing of Asian units of measure. I deduce that they use decimal units, but are their units unique in size to that part of the world? What do you know of the origin of the “Sashigane”? I have been told in conversation that the Chinese were first to use it over 3000 years ago. Is that accurate information as far as you know?

It’s been great talking to you and I wish you the best of luck on your journey. Stay in touch. We’ll be here.

Chris Hall
05-03-2005, 02:28 PM

your thoughtful comments are highly appreciated. I don't have much time at the moment, but I can respond to your question about the sashigane.
The sashigane derives from an older Chinese tool, which i've never seen but was like a square with an extra right-angled arm on it. That tool gives direct rise to a sino-Japanese written character read in Japanese as 'takumi', which means 'craftsmanship'.
The original sashigane were made of bamboo, and now are stainless steel. The traditional Japanese measuring system, 'shakkan-ho', stems from a Chinese antecedent, and, like the English inch, varied somewhat in early architectural work, only becoming standardized in Japan after the unification of the country and the commencement of the Edo Period.
The basic unit is the shaku, which is 11.93", or very close to an English foot. The shaku is divided into 10 units, each of which are called 'sun'. The character depicting this measurement of 'sun' is a pictograph of a hand with a tickmark on the wrist - the distance of the pulse point actually, at 3.03 cm below the fold of the wrist.
Sun are further divided into 10 'bu', and those are further divided into 10 'rin'. Going the other way, 6 shaku make for another unit called 'ken', which happens to be a standard length for a tatami mat, the primary organizing module for Japanese rooms. An excellent commentary on shakkan-ho can be found in the Engel book mentioned earlier in this thread.
Sashigane have a sq.rt. 2 scale on the back to simplify regular hip and valley layout. Some also have a pi scale so as to convert diameter measures of round objects into circumference, and a depth scale on the bottom of the long arm.
There are sashigane in Japan in both shaku-sun-bu, and in metric. There was one produced for the western market in inches, but regretably in base 12, with 16ths and 32nd. While this is normal for western roof work, Japanese roof framing uses base 10, so that sashigane was pretty much useless except for basic measuring operations.
When I got a chance a few years ago to design a new square for Shinwa, I was thrilled to finally redesign the square to make it useful in inch-scale. Thus it is in base 10, with divisions of 1/20th of an inch. This eliminates the need to convert from decimal, and go straight to the square from your calculator.
Speaking of calculation, it is possible to use the square all by itself to determine both square- and cubic-roots!

Chris Hall
05-04-2005, 03:17 PM
I realized this morning when I woke up that I had made a mistake yesterday when ascribing the Japanese character for 'takumi' to a pictograph of a early type of Chinese framing square. In fact, that character derives from a pictograph of an adze. The character in Japanese that derives from the square mean's 'craft' or 'industry', and looks somewhat similar to the one read 'takumi'.
I realize that this point will be irrelevant or of no interest to most readers, but I wanted to be accurate.