I have written a code, and I need a little feed back..

I want to make the Subscript distributive across (1) addition and (2) multiplication to mimic short-hand for added vectors. For example, in $(p+2k)_i=p_i+2k_i$ the LHS is short-hand for RHS, where the index is not allowed to be on the coefficient 2. Finally, I need exception, where the numeric quantity 0 should be allowed to take an index: that is, $0_i$ should not go to $0$.

My code is:


Subscript[Plus[a_, b___], idx_] := 
  Subscript[a, idx] + Subscript[b, idx];

Subscript[Times[a_, b___], idx_] := 
  Times[Subscript[a, idx], Subscript[b, idx]];

Subscript[a_?NumberQ, idx_] /; a != 0 := a;

What I did was to define Subscript as functions, which is already built-in Symbol in Mathematica. The code seems to really work, but I don't know if I deserve a pat-on-the-back for being clever, or be reprimanded for overwriting a built-in Symbol (or for any other bad coding behavior).

Also, is there any obvious way anyone can see this piece of code going wrong? Thanks!!

  • 1
    $\begingroup$ I would use ClearAll[Subscript] in case Attributes has previously been set to something undesirable. Otherwise, there's nothing wrong with defining a function like Subscript, because it has no built-in meaning. However, I don't know what the real intent is, so it's hard to write anything other than a comment. E.g., what about dot products, or Subscript[a + b c,i]? That result doesn't look like something you normally want. $\endgroup$
    – Jens
    Jan 27, 2013 at 21:25
  • $\begingroup$ @Jens Ah! Very good point about dot products. I suppose that this is restricted to sums of vectors, and products of numbers and vectors. But not square of vectors. Thank you for feedback. $\endgroup$
    – QuantumDot
    Jan 27, 2013 at 22:12

1 Answer 1

  • In my opinion Subscript is a formatting function and should not be occupied with this kind of definition. With such a definition in place you will have to rely on SubscriptBox any time you want a simple subscript without these rules.

  • Your definitions as written insert a raw sequence into Subscript that I do not believe you want. I shall assume that $(a+b+c)_7$ should be expanded to $a_7+b_7+c_7$ and not $a_7+b_{c,7}$

To correct these I would define:

ss[a_ + b___, idx_] := ss[a, idx] + ss[+b, idx];

ss[a_ * b___, idx_] := ss[a, idx] * ss[1 b, idx];

ss[a_?NumberQ, idx_] /; a != 0 := a;

Format[x : ss[args__]] := Interpretation[Subscript[args], x]


ss[q + r + s + x y z, 7]

$q_7+r_7+s_7+x_7 y_7 z_7$

% // InputForm
ss[q, 7] + ss[r, 7] + ss[s, 7] + ss[x, 7]*ss[y, 7]*ss[z, 7]

Jens comments that attaching a definition to Subscript in the way you did is perfectly acceptable. If you choose that option then you will probably want to define a new subscript function for the purposes of pure formatting. For example:

MakeBoxes[realSubscript[a_, b__], fmt_] := 
  SubscriptBox[a, RowBox @ Riffle[{b}, ","]]

realSubscript[1, 2]
  • $\begingroup$ The tutorial doesn't agree with you. An operator without built-in meaning can be a very convenient way to define a new function that automatically comes with a notational form. The latter doesn't mean that it's only a formatting construct. $\endgroup$
    – Jens
    Jan 28, 2013 at 3:23
  • $\begingroup$ @Jens That's fine. I don't always agree with the documentation or the popular opinion. I would (and do) use other operators directly. However I use subscript in labels, etc. often enough that I would not want to have to define a new subscript function though I suppose that's an option for others. $\endgroup$
    – Mr.Wizard
    Jan 28, 2013 at 3:41
  • $\begingroup$ Sure - you marked the statement as an opinion - I just wanted to mention the alternative opinion. $\endgroup$
    – Jens
    Jan 28, 2013 at 3:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.