"Simon St.Laurent" wrote:
>
> (james anderson) writes:
> >> While that is compelling, it's probably not acceptable given that
> >> you may have multiple consumers of your data who may or may not be
> >> machines.
> >
> >but how many of those non-machines will not have their access mediated
> >by one?
>
> Depends on what you mean by "mediated". While I do occasionally work
> with XML on paper or blackboards, I generally work with it in text
> editors. I occasionally (10% ?) edit XML indirectly through an
> application, which is a heavier form of mediation. I don't think I ever
> work with XML expecting _never_ to have to look at the markup.
>
> In any case, I design my own tools around the expectation that mediation
> may be available but will not be necessary for interpretation.
>
on the other hand, i can't put my finger on the last time i seriously tried to
interpret a stack trace without a symbol table. or rather without some machine
doing the interpretation for me. and even in the days when i had to, it never
would have occurred to me to expect to find my comments in the machine code.
...
