# Best practices: typesetting multidimensional functions with multiple arguments [closed]

Consider, as a prototyical example, the following code:

$F\!\begin{pmatrix}a_{1}\\ a_{2}\\ a_{3}\\ a_{4} \end{pmatrix}=\begin{pmatrix}b_{1}\\ b_{2}\\ b_{3} \end{pmatrix}$

which produces

Notice, for instance, the small negative space that I inserted after F to make the output look good.

Can you please tell me, if this you pass of as a typographically "correct" (or at least generally accepted) solution to typesetting such functions ?

Are there any improvements that can be made ?

Side question: Is there any rule, when it is better (say, after a certain number of arguments) to switch to displaying the parenthesis in a horizontal format, i.e. "row vectors" ?

• The thin space between F and the matrix is a consequence of pmatrix being surrounded by \left( and \right). This space is good when two matrices are next to each other, but in this case it might well be removed and \! is the right solution. Of course, if you use the method suggested for the macro definition in the answer to your recent question, you can insert \! as part of the definition itself and get it automatically. – egreg Sep 25 '16 at 20:12
• @egreg Yes, exactly that was the whole I asked that question ;) for any ulterior change, when someone points me to some fine detail, won't make me rewrite the whole document, but just change the definition of that command. Though the space between the = sign does look pretty big.... – l7ll7 Sep 25 '16 at 20:49
• whether it's possible to switch to row vectors, and whether you can just switch or need to mark each transposed matrix with ^T depends a lot on the mathematical context unlike the space adjustments that's not really a typographic question. – David Carlisle Sep 25 '16 at 20:56
• That probably makes the question off-topic. We answer questions about how to do it. Whether to do it is somebody else's problem ;). (Maths SE or the typography/graphics design SE.) – cfr Nov 14 '16 at 17:12
• Typesetting like that tends to be cumbersome. Especially if it's repeated everywhere. Instead I'd say something like: given [\mathbf a\in \mathbb R^4,\mathbf b\in\mathbb R^3: \mathbf F(\mathbf a)=\mathbf b]. – Klaas van Aarsen Nov 14 '16 at 17:47