Whatever an architect does, he always does it in space, in 3 dimensions, right?

Architecture is a spatial profession. Elevations and sections, etc. are only translations of space onto a 2D plane. The code, or script, I’ve writing about earlier – are only necessities for generating space. Or are they? May-be they actually are “the” architecture?

One could argue that drawings are not the consequence of space, but its generators. I guess that is how it has been frequently understood before. Architects drew first and with that they defined the 3D space. Till this very day this practice flourishes: on the desk beside mine in the studio I’m in, there is a guy, who continuously makes elevations, plans and sections as distinct entities and then kind of blends them together into a spatial 3D model – in his case an axonometric projection. A very exciting way to make space, I should say, and demanding great skill. Translating your 2D construct into 3D, you find yourself “discovering” the space you’ve created.

However, there is also a different way of understanding the 2D-3D relation in Architectural design. This relation could be perceived as a reverse of the previous:

3D space -> drawing,

as opposed to drawing -> 3D space.

In this case a drawing – is a consequence of space, a “necessary evil”, a translation of a far more superior 3D into 2D for purely utilitarian reasons – to facilitate the manufacturing process, to make readable for the contractor, etc. The real thing is always the three dimensional space. That is why many people usually make a physical model first and then “digitize” it or translate it into drawings. Or, in case of 3D modeling, produce a 3D model first and then slice the model into pieces to produce drawings. This approach has incredibly grown in popularity with introduction of complex geometries, which are simply impossible (or possible, but ridiculously hard) to draw “by hand”.

From this dichotomy a legitimate question arises:  which way is the right one? Which one is the best?

I, personally, do not think that any one way is superior to another – both are completely capable of producing exciting spatial experiences. Yes, one tends to be more rectilinear, but no one said there’s something wrong with that. And in the end, in Architecture, what matters is not how you designed it, but what you designed and whether it’s good. So:

3D -> 2D

or 2D -> 3D

or 0D -> 3D (as with “blind” scripting I’ve described before)

You choose according to your liking.