Skip to content

Architecture Misfits #38: The Unknown Architect

Post date:

Every now and then, instead of taking my usual route to work, I’d head east on the E44 towards the oasis town of Hatta and hook up with the E611, the main inter-emirate freight-haulage route and exit at the Al Badi Interchange that’s now one fourth orbital.

On Maliha Road now going in the direction of Sharjah on the left there’s a strip of light industrial and commercial and on the right a strip of curious mixed-use buildings with residential beyond. I’d never really looked or thought much about any of it until I noticed this next building I’ve circled it on the aerial photo below.

Having noticed one, I started to look at the others of which there are about 40, with perhaps 15 in various stages of construction. Along these first and second side roads there’s probably plots for at least 200 more.

They’re all of a very precisely defined type with retail units on the ground floor and studio or small apartments on the upper floor. Of the ten or so already occupied, there’s one or two restaurants, a couple of convenience stores and the remainder are building supply outlets, including three for National Paints, curiously. I imagine the others will fill up with outlets for door furniture and other building supplies that don’t require huge inventory. This is how the area looks now.

Construction isn’t complex.

The only two sub-types are mid-terrace and end-of-terrace. All buildings are at least twice as deep as they are wide so nearly all mid-terraces have a side upper terrace to allow daylighting to apartments opening onto it. Some have this first floor terrace on the right side.

And some have it on the left. It’s uncommon to see two buildings the same but two of the following six are identical and a third almost. That there are many different contractors suggests these buildings are an opportunity for small-scale investment.

Which side the terrace is on doesn’t have anything to do with aspect as the terrace of one building can abut that of its neighbour.

The sides of mid-block buildings are blank and the residential entrance can be either front or back. In many cases the buildings are mirrored front to back structurally but not functionally.

Larger developments usually comprise repeats of mid-terrace buildings. Below you can see the front-back symmetry and how the side terrace is allocated.

My initial example was a rare case of a monolithic triple development, with the remainder of the block taken up by two repeats.

This next example is the rear of a building without terraces. Each of the three units has two or three apartments front and back, with a corridor running along the length. This arrangement can be used to configure developments of any length.

This is one of only two paired developments that went for mirror symmetry.

End-of-terrace buildings have more surface area and the absence of terraces confirms their primary role to supply light and ventilation. These apartments have no outdoor space but the recessed balconies front and back shorten the internal corridor. As before, there are left-end buildings

and right-end buildings.

The absence of the end balcony in this next four suggests a corridor running across the width instead of along the inner side.

This brings us to the corner problem. I think what’s happened is that the end site corner truncation of about 1.5m was decided when the subdivision was laid out, and the uniform setback for the shopfront decided independently some time later. Below is my understanding of the problem. I’ve drawn the main grid at 3m x 3m and distance A at 2.00 metres. A ≠ 3.00 metres and so we have this problem of the unsupported corner.

These corner buildings are four and 2/3 bays wide but both corners of the truncation don’t fall on column positions, or at least they won’t if it’s 45° and you wish to maximize rentable area. This leads to solutions such as this one, where the corners of the truncation are supported by columns but only one of those columns is on the structural grid followed by the rest of the building. Such conflicts between where columns are and where they need to be aren’t unique.

The most famous is the courtyard at Florence’s Palazzo Strozzi, built 1489-1538. The problem was that the corners look visually weak if a single column is placed where the grid lines intersect at the corners. What to do?

Here’s what Bramante did at the cloisters of Santa Maria della Pace in Rome, 1500-1504. Instead of columns, he went for repeated openings that created more wall area to define the sides of the courtyard more strongly. The text frieze also helps, as do the pilasters added to the short runs of wall between openings. There’s no pilaster on the corner but the 45° infill between the walls hints at one while also working to make the lower part of the corner more massive.

At Palazzo Strozzi, architect Benedetto da Maiano had faith in the rules and simply applied them. At Santa Maria della Pace, Bramante didn’t like what the rules produced and so he adjusted as he saw fit. At Seagram Building, Ludwig Mies did both, producing an over-publicised external corner detail to indicate his sensitivity to historic concerns, while the internal corner with the greater claim to history was allowed to happen as the rules dictated.

The designer of this building also did both, following the rules where it was expedient to do so and, where they didn’t work, overcompensating in order to show that a design effort had been made. I don’t see it as example of complexity and contradiction. It’s more like a workaround that doesn’t work. I say that mostly because the height of the lintel on the corner is determined by the corbels springing from the column on the right.

It might be worth a try removing that extra column and adding more corbels to lift the lintel to full height, and then offsetting the window to the apparent new midpoint. It’s better than some other “responses” to the same problem.

As it stands, the lintel is neither high nor low and not exactly a transition either. It draws attention to the solution without actually solving the problem. It illustrates that hoary old maxim “If you can’t solve a problem then make a feature out of it.” The most common way this problem has been solved is to just put an extra column under the extra corner and leave it at that. Job done.

A more elegant solution is to extend the side wall even though that difference in lintel heights isn’t going to go away. It’s strange in this next example how the shape of the corner column and wall supports refuse to acknowledge the truncation and the reason they’re there.

In this next, the building on the left goes for the simplest solution but the one on the right purposely ignores the site truncation and so sacrifices some internal area, only to claw some of it back with a cantilever that partially reinstates the truncation. It’s a snapshot of the design process still unresolved. One can imagine this building still trying to work out what it wants to do. [If you want to really overthink it, it’s a new dimension to dynamism and a new take on Deconstruction. Is this building frozen while morphing towards some perfect state or away from one? etc. ]

Here’s a sunnier view of the same exciting corner.

One new problem is that if the building truncation exactly follows the line of the site truncation, then the 90° corner above must either overhang the site boundary and/or the cornice above it must exceed any allowance it might have been granted.

I expect the designer received a slap on the wrist for one or the other because two buildings up is a sister building that doesn’t have this problem. Instead, this pseudo-solution produces more of the 45° wall it was supposed to downplay. It’s interesting how every element and surface of this corner is firefighting some localized problem brought about by this conflict between the grid dimensions and the size of the corner truncation.

That supported “keystone” is another example of solving a problem in a way that draws attention to the problem which, in this case, was to stop the wall dividing the balcony from destroying the visual integrity of the arch.

All this comes about because of the desire to avoid an unsupported corner. These next three buildings don’t care. You decide if it’s worth worrying about.

These three are the closet any end-of-terrace building comes to ignoring this 45° site condition. Not one building altered the grid to put a line of columns one meter or so in front of the shopfront although, admittedly, 1) this would work against the purpose of shopfronts and 2) may be seen as a willful misinterpretation of “setback”. And not one building connected the last columns on the facade and shopfront grid lines to produce a corner wall not at 45°. This would entail some loss of area but I suspect it’s a cultural thing where one quarter of an octagon is still better than no octagon at all.

This one’s my favourite. Is it beautiful – who knows? If you agree, as I do, with Subjectivist philosopher David Hume’s statement that “Beauty is no quality in things themselves. It exists only in the mind that contemplates them” then something considered beautiful means nothing more than that it being suited to evoking certain responses in us. Here are mine.

  • The lintel height is unified by making the front one lower. I like this.
  • The side wall is extended to the corner of the truncation, concealing the corner column that’s not on the grid and which is the reference for the spacing of the side windows. Rules set up by the column positions have been used to solve this single column incongruity as simply as possible without ignoring it, concealing it, or making it into a feature.
  • The columns are not minimal width, and regulate the upper floor as well. Good thinking.
  • The end column on the right is the same width as the front surface of the corner column on the left and evidence somebody thought about this. Admirable.
  • Each of the windows and their positions has been thought about. On the front facade, the leftmost upper window isn’t as wide as the others and is moved slightly to the right to make the overall spacing more uniform. More evidence of thought.
  • The amount of ornament is sufficient to emphasize dignity but sparse enough to avoid ostentation or accusations of pretence. This is rare. Stunningly so.
  • The rectangular windows on the side elevation are mostly in line with one side of the openings below and the curved windows equally spaced between them inasmuch as the column positions and interior walls permit. This is no small problem. When windows are of different widths, placing each on the vertical axis of the openings below will produce uneven spacing, but placing each window midway along the wall of the room it lights will also produce an uneven spacing on the exterior. A balance has to be found between how the window looks on the inside, and how it looks on the outside. Column positions will restrict exterior options so it’s not a totally free call. Asnago Vender were masters of window positioning. So too were the designers and later re-designers of Venice’s Doge’s Palace. The circular windows show how expedient deviations to a pattern are perceived as regularities if the pattern is sufficiently strong.
  • The only thing I don’t understand is why the middle extension of the parapet on the front facade isn’t in line with the windows. I only noticed because of those parapet moldings (that could easily have had their positions fudged to disguise this offset if it had been some construction error). However, the same thing occurs on the side and rear elevations but for reasons I still can’t work out.
  • The truncation hasn’t been made into a feature as it has for every other corner building although leaving it blank is distinctive in its own way.

I could always phone the number of the agent or owner in the sign on the front of the building, ask who its architect was and say how much I like it. I may yet, but what’s more important is that this building is there and to my mind at least far better than it needed to be, than anyone was expecting or in fact anyone even appreciates. It won’t change the world but it’s made one tiny corner of it better than it was before. Its architect is not well known and may never be but this is not tragic or sad. Rather, it makes me happy to see people enjoying what they do.

• • • 

Unknown architects everywhere,
doing your best with whatever jobs come your way,

misfits salutes you!

• • • 



  • Hi Graham, greetings from Mexico. I have been following your blog for a while now. I’m currently reading Matt Ridley’s book How Innovation Works and it somehow reminds me of the thoughts and views expressed in this blog. If you haven’t read it yet, I definitely recommend it.

    • says:

      Thanks Gerardo – I enjoy receiving and following up on recommendations like this. I’ll get back yo you. Graham.

  • An age old question. Best solution is to ignore the question?
    With the amount of road pavement in front and to the side, the truncation appears superfluous?
    I like the notion of the continuous veranda or arcade on the principal street facade. Excellent idea. Not sure if it is owned by the retailer or the public or shared.
    In many cases it seems to not turn the corner? Seems a pity.
    The side terraces are also a curiosity. Are they a traditional device?

    • says:

      I’d agree the truncation isn’t really necessary – it’s just something to do on a corner. As for the arcade, I suspect they’re not rented by the retailers but they probably have limited rights to use it for display or tables during operating hours. And yes, it is a pity most don’t turn the corner but it’s easy to guess why. On the end-of-row units that extra 2m width represents significant retail area which is of course monetized in all the middle-of-row units. Directly above it are those terraces which use this 2m wide strip as open space enabling perhaps another two apartments along the length of one side of the building, assuming two each front and back. I suspect it’s no more than a light and ventilation thing because the end-of-row buildings don’t have them. What fascinates me about these little buildings is how little variation there is. The rules were set and then followed and certain solutions arrived at more frequently than others. I’m still trying to find some upper floor plans.

      • Interesting. Not quite Victorian terrace Housing but sort of.
        I assumed that the upper levels are residences but I am curious about access to the terrace and if the units are multi levelled or ?
        It seems like an appropriate solution all round.

      • says:

        I wondered about that too. This photo shows how the apartment or apartments with windows to the front (or rear) might not have access to the terrace, suggesting its primary reason is to provide light and air. If it’s used as outdoor amenity space at all, it’ll probably be for laundry drying.

        The fact it’s there is probably a case of a problem being solved in the simplest way possible. That problem is most likely something like getting the maximum number of apartments with some minimum daylighting requirement, if for some the only window will be facing a 3m-high blank wall. I also think it’s an appropriate solution, and one that was arrived at amazingly quickly.The best thing to do was decided almost immediately. I think this is why the cosmetic differences are so observable. When driving around the area and looking at all these buildings, I feel like I’m looking at a Design I submission. Or wish I was. This little building and its few simple rules would make an excellent early design project.