Mezzanine problems in Cornell’s proposed Fine Arts Library

Links to all my writings and blog posts about Cornell’s proposed Fine Arts Library can be found here.

The Fine Arts Library proposed for Rand Hall at Cornell University has two stories, each with a mezzanine floor, for a total of four levels all linked by a single continuous vertical opening. Vertical openings in buildings are—with some important exceptions—prohibited by Building Codes because they present a serious fire and smoke hazard if not properly designed or protected by a continuous and fire-rated shaft enclosure. The most common exception to the general prohibition of vertical openings states that any two stories may have a vertical opening between them. This much is clear. However, when mezzanines are introduced, things get more complicated and additional exceptions to the general prohibition against vertical openings must be analyzed.

The 2003 and 2007 versions of the New York State Building Codes—the first that were based on the International Building Code (IBC)—state in Section 505.1 that: “A mezzanine… shall be considered a portion of the floor below.” This is important because the exception allowing two-story vertical openings (in Section 707.2 of the 2010 NY State Code or Section 712.1.9 of the 2015 NY State Code, for example) specifically allows such floor openings to connect two stories, but only as long as they also are separated from other floor openings. Because mezzanines were previously defined as being part of the floor below, they were not considered as separate floors, and therefore didn’t count as additional floor openings under the terms of this exception. In other words, the four interconnected floors in the Fine Arts Library proposal might have been acceptable under these older Codes (assuming that all other requirements governing maximum mezzanine size were satisfied). However, in all subsequent versions of the New York State Building Code, including those from 2010 through 2015, a single word in Section 505.1 (or 505.2 in later versions) was changed, so that a mezzanine is now considered a portion—not of the floor below—but of the story below.

With this seemingly innocent change—from floor to story—the exception in Section 707.2 of the 2010 Code (or Section 712.1.9 of the 2015 Code) allowing floor openings connecting no more than two stories as long as they are also separated from other floor openings now has a very different meaning. With mezzanine floors no longer included as part of the floor below, but still considered as independent floors, the requirement that any vertical opening linking two stories be separated from any other floors is no longer satisfied if mezzanines are present. It is no longer possible to ignore the mezzanine floor levels by claiming that they are part of the floors below them.

There are only two other potentially viable options for designing “vertical openings” for a new Fine Arts Library in Rand Hall, per Section 712.1 in the 2015 NY State Building Code. The first such option—having a vertical opening between a mezzanine and the floor below—clearly doesn’t work for the Fine Arts Library proposal, since only a single mezzanine and floor can be linked together using this option. The only other viable strategy is to design the vertical opening as an atrium.



The mathematics of the symbolic skyscraper

In Delirious New York1, Rem Koolhaas describes a crisis in symbolic expression allegedly caused by the development of the skyscraper. First, he questions whether something “filled, from top to bottom, by business” can even be considered as a monument with symbolic content, other than as an Automonument with an “empty” symbolism that “does not represent an abstract ideal, an institution of exceptional importance, a three-dimensional, readable articulation of a social hierarchy, a memorial…”

The idea that buildings for church and state are legitimately monumental (hence having symbolic meaning) but buildings for business are not is, on the face of it, absurd. The attacks on the World Trade Center Towers in 1993 and 2001 provide some evidence that these skyscrapers for business2 (i.e., “world trade”) were symbolic of economic activity and economic power, and that precisely this symbolism was a factor in their selection as targets of terrorism. But even ordinary business towers have meaning (hence, act as symbols): the “corporate officers and the army of clerks they employed determined the meaning and usage of the tall office building” according to Roberta Moudry in her introduction to The American Skyscraper.3 Aaron Betsky is even more emphatic, stating that “Of all the buildings human beings make, towers are the most symbolic” and specifying that they “stand for the wealth and power of the commissioning agency, whether it is a developer, a private entity such as a corporation, or a (quasi-) governmental agency” (in Asia and elsewhere) or “are symbols for the transformation of Manhattan into the haven of choice for the global one percent.”4 That Betsky’s towers include both commercial and residential types does not matter in this context. And, of course, one can always cite Louis Sullivan for paeans to the symbolic grandeur of the skyscraper for business: “that the problem of the tall office building is one of the most stupendous, one of the most magnificent opportunities that the Lord Of Nature in His beneficence has ever offered to the proud spirit of man.”5

But aside from this assertion that the meaning of office towers is somehow second-class and unsuitable for monumentality or symbolism, Koolhaas makes a second claim that is equally absurd: that the height of skyscrapers makes it increasingly difficult to express internal functionality on outside surfaces, something allegedly required within the tradition of the “honest” facade in modern Western architectural culture. Koolhaas states that “mathematically, the interior volume of the three-dimensional objects increases in cubed leaps and the containing envelope only by squared increments: less and less surface has to represent more and more interior activity.”

Figure 1

But as is clear from Figure 1, the ratio of volume and surface remain unchanged as a building gets taller. The mathematical truth is that making a building higher does not change the relationship of volume to surface; only increasing plan dimensions does that (Figure 2).

Figure 2

Thus this alleged crisis (“Beyond a certain critical mass the relationship is stressed beyond the breaking point…”) has nothing to do with the skyscraper or the tower — i.e., with height — but only with the size of the block. And block dimensions in Manhattan were already fixed in size with the Commissioners’ Plan of 1811.


1. Rem Koolhaas, Delirious New York: A Retroactive Manifesto for Manhattan, Oxford University Press (New York: 1978), p.81-82.

2. Yes, it’s true that some quasi-governmental organizations, such as the Port Authority of New York and New Jersey, were tenants of the World Trade Center, but the bulk of space was rented to business: see “List of tenants in One World Trade Center,” Wikipedia (accessed 2/14/17)

3. Roberta Moudry, ed., “Introduction,” The American Skyscraper: Cultural Histories, Cambridge University Press (Cambridge, New York, etc.: 2005), emphasis added

4. Aaron Betsky, “Symbolism of Skyscrapers: The Meaning of High-Rises Around the World,” (The Journal of the American Institute of Architecture), July 23, 2014 (accessed 2/14/17)

5. Louis Sullivan, “The Tall Office Building Artistically Considered,” Lippencott’s Monthly Magazine, March 1896, p.406 (text available here)

Cornell’s proposed Fine Arts Library at 50% design development

UPDATED Feb. 13, 2017 (see comments below).

I’ve been out of the country since August 2016, so I was not able, until now, to examine the 50% Design Development (DD) drawings for the proposed Fine Arts Library that have been made available for viewing in the Dean’s office. This post, therefore, may well be old news. However, having already noted numerous problems with the schematic design proposal, I was curious to see if any or all of them had been resolved. As far as I can tell, the proposal still has many problems, some old and some new. In particular, it appears that the fire wall shown in the schematic plans has been eliminated (replaced with various “fire barriers”) so that the advantages of considering Rand Hall as a separate building (only possible with a fire wall) have apparently been discarded. I’ve sent out an email today with a bunch of questions for the contact person at Cornell Facilities, reprinted below, and will update this post if I get any answers.


1. The Building Department Notes in the 50% DD set list the 2010 NYS Building Code and 2010 NYS Existing Building Code as governing Code documents. Will the drawings be changed to reflect the 2015 NYS Building Code and 2015 NYS Existing Building Code now in effect?

2. There is no comprehensive Code analysis in the 50% DD set that explains the basis of the design. For example, the schematic design proposed a fire wall between Milstein Hall and Rand Hall, allowing Rand Hall to be considered as a separate building. There is no fire wall shown in the 50% DD set (except for “mobile fire wall” notations on one of the elevations — see Question #6 below); therefore Rand-Milstein-Sibley Hall must be considered as a single building with a single construction type. This construction type is V-B, based on the 3rd-floor wooden wall construction of Sibley Hall, and not II-B as noted in the DD set.

A Code variance granted in November, 2013 acknowledges the fact that the existing library as designed and constructed on the third floor of Rand Hall (or a proposed, but not yet designed, library on the second and third floors of Rand Hall) would be noncompliant under the 2010 NYS Building Code, and requests that two specific 2010 Code sections be waived. As a result of the variance hearing, certain requirements in Sections 503.1 and 504.2 of the 2010 NYS Building Code were waived, allowing an increase in the allowable floor area of the second floor from about 22,500 square feet to 70,000 square feet and permitting the library to exceed the second-story limit stipulated in the Code (based on its V-B construction type and A-3 occupancy class) and to occupy the third floor of the “Rand Hall” portion of the combined building. A third Code section waiver, not originally requested by Cornell but suggested by the Hearing Board, waives portions of Section 715.1 of the 2010 NYS Building Code so that the requirement for opening protectives (for the windows that penetrate the fire barriers between Milstein, Rand, and Sibley Halls) no longer needs to be applied in considering the Code compliance of the existing library.

This variance clearly allows the current third-floor library to remain in Rand Hall. It does not, however, change the construction type of the combined buildings from V-B to II-B. It should be emphasized that without a fire wall, Rand Hall is not a separate building, even with the variance, and its construction type remains V-B.

On what basis is the construction type for this project listed as II-B?

3. A Code variance specifies what can be constructed, not merely by enumerating practices that are deemed allowable for a particular project, but rather by waiving specific Code sections that would otherwise render the construction noncompliant. The Code variance granted Cornell in November 2013 waives three Code sections in the 2010 NYS Building Code, allowing the current third-floor library to remain. However, those Code sections are no longer applicable to new construction or renovation, since a new 2015 Building Code has taken effect. The variance granted in 2013 for the current library did not waive any Code sections in the 2015 NYS Building Code. Do you believe that a new library built under the 2015 NYS Building Code (and 2015 NYS Existing Building Code) can waive floor area, story height limits, and window protective requirements in the new Code on the basis of a waiver of three Code sections that were granted with respect to the existing library under the 2010 Building Code?

4. Since the building construction type of V-B is neither Type I or Type II, the maximum aggregate area of a mezzanine is 1/3 of the floor area of the room in which it is located. On the basis of what Sections in the 2015 NYS Building Code can the mezzanines shown in the 50% DD set exceed this 1/3 floor area limit?

5. Section 712.1.9 (Two-story openings) of the 2015 New York State Building Code requires that all floor openings must be “separated from floor openings and air transfer openings serving other floors by construction conforming to required shaft enclosures.” Mezzanines, while not counted as “stories,” are still counted as “floors”; therefore openings connecting any more than two such floors are not permitted unless protected by a shaft enclosure, or designed as atriums. How are floor openings connecting all four floors (not stories) in the Fine Arts Library permitted under the 2015 Building Code?

6. There are two notations on the West Elevation in the 50% DD set that say something like “mobile fire wall” (the font is so small that, even with my reading glasses on, I can’t make out the exact words) and point to the north-west and south-west corners of Rand Hall. What do these notations mean? What is a “mobile fire wall”?

7. A sectional drawing in the 50% DD set shows a dimension of 2’-3” below the hanging 2nd-floor stack area, but seems to show protruding objects on the south side of this hanging floor that are higher than 2’-3” and therefore in violation of the 2015 NYS Building Code and ADA. Can you explain what these protruding objects are (see sketch below) and why you believe that they are Code-compliant?

UPDATE Feb. 13, 2017: I received a brief email response from the Project Manager/Sr. Engineer at Cornell Facilities: “Thanks very much for your thoughts and input and I will forward your email to our design team.  It appears that you are reviewing a 50% DD set, which has evolved considerably.  Once the DD phase is complete there will be new drawings that may address many of your questions.” In other words, it looks like I won’t have any definitive answers to these questions until the 100% Design Development drawings are released.

Links to all my writings and posts concerning the Fine Arts Library proposal can be found here.

Koolhaas delivers “junk-space” to Cornell, in his own words…

“There are no walls, only partitions, shimmering membranes frequently covered in mirror or gold.”


“Structure groans invisibly underneath decoration, or worse, has become ornamental…”


“…huge beams deliver cyclopic burdens to unsuspecting destinations…”

Rem Koolhaas, “junk-space,” Brendan McGetrick, ed., Content, Taschen (Köln, London, etc., 2004), p. 163. Images are of Milstein Hall, Cornell University, designed by OMA.

The NY Times Interprets Architectural History

“He became fascinated by architects like Louis Sullivan, who stripped the veneer off their buildings and let the strength of their construction shine through.”

Huh? Louis Sullivan? Stripped the veneer off his buildings?

Louis Sullivan's Prudential Guaranty Building, in downtown Buffalo, New York (photo by TomFawls, Wikipedia)

Louis Sullivan’s Prudential Guaranty Building, in downtown Buffalo, New York (photo by TomFawls, Wikipedia)

Quote is from: Jon Mooallem, “One Man’s Quest to Change the Way We Die,” The New York Times Magazine, Jan. 3, 2017.

Teaching at Tianjin University School of Architecture

I received a U.S. Fulbright Scholar grant to teach at Tianjin University School of Architecture for the fall 2016 semester. We leave in late August and return in late January 2017. The plan is to teach two courses about architectural practice: one for undergraduates dealing with building technology and the other for graduate students dealing with politics and sustainability. I’m looking forward to the adventure, with just a bit of trepidation about the pollution.

This is me PhotoShopped into the Beijing marathon; original image from

This is me PhotoShopped into the Beijing marathon; original image from

Exit signs in E. Sibley Hall at Cornell

Exit signs are supposed to clearly show you the way out of a building. The Building Code of New York State puts it this way: “Where required: …The path of egress travel to exits and within exits shall be clearly marked by readily visible exit signs to clearly indicate the direction of egress travel in cases where the exit or the path of egress travel is not immediately visible to the occupants…” With all the best intentions, Cornell University replaced some old and ordinary doors and walls within the main egress stairwell of its architecture building (E. Sibley Hall) with fire-rated glazing and installed a bunch of illuminated exit signs to show the way out.

Views of exit signs from inside the E. Sibley Hall egress stair

Views of exit signs from inside the E. Sibley Hall egress stair

Unfortunately, as can be seen in these recent photos, the transparency of the doors and walls violates the spirit, if not the letter, of the Code. This happens because as one reaches the second floor in the stair (image “b” on the right), one sees an exit sign apparently pointing to an exit outside the stair, just like the sign that one sees on the first floor (image “a” on the left). However, while the sign on the first floor actually points to the exit discharge, the sign on the second floor points away from the exit discharge. In fact, this second-floor sign in intended to get you into the stair from the outside, not to induce you to leave the stair. However, because people using the stair shaft can now see exit signs outside the stair that were not meant to be seen from within the stair itself, what was a simple path of egress has become confusing and therefore dangerous.

Steel truss design

I’ve written a paper (not yet online) about graphical statics. In order to demonstrate that the “form-finding” objectives of such techniques are often superficial and flawed, I needed to provide a case study of an actual structural design problem, using real materials and real design methods, that accounts for things (like the buckling of bars in compression, or shear lag and effective net area of bars in tension, or deflection issues) that graphical statics ignores. So I created an enormous spreadsheet to find the optimal (lightest) steel double-angle Pratt or Warren truss for any given span, spacing, and loading condition. The spreadsheet actually designs 105 different trusses (with aspect ratios ranging from 2:1 to 16:1; and with from 2 to 14 “panels,” as shown in Figure 1) in order to find the optimal combination of aspect ratio and panel geometry for a given span, spacing, and loads.

Figure 1. Trusses are designed with from 2 to 14 panels

Figure 1. Trusses are designed with from 2 to 14 panels

Well, the spreadsheet had well over 1,000,000 cells and couldn’t be converted to an online calculator using the software I have. Therefore, I made a smaller version of the same spreadsheet — this one only designs a single truss (rather than 105 versions) and only considers whatever aspect ratio and panel geometry have been selected. However, this calculator can also be used to find the lightest truss, but only by trying out numerous geometries while keeping track of the truss weight.