>From: Per Bothner <xxxxxx@bothner.com>
Subject: Re: new srfi-118 draft
Date: Wed, 18 Feb 2015 15:44:10 -0800

> I could add to string-append!:
>
>   There is no requirement for this procedure to execute in constant
>   time.
>   In some implementations it execute in amortized (i.e. average-cost)
>   constant time.

This is what I'm afraid of... people tends to assume something
like this, unless it is explicitly noted that some implementation
execute it in O(n), and there's no performance advantage from
string-append.