Posted by mbstevens on 12/09/07 06:20

dorayme wrote:

> The remark by mb meant something to him and to me. It being an
> aside, it was not going to be explored in detail. It was accurate
> enough for the purpose at hand, OT as it may be. But your ears
> are finely programmed to catch anything said that, taken quite
> literally, in a context of your own choosing, is not strictly
> correct.


However he wants to spend his
_valuable_ time is OK with me. :)

I did gloss over that the validator will do limited things like
warning you of mistakes that are 'semantic'
-- like an h2 appearing before there is an h1.

As you point out, it was a side issue that had no direct
bearing on the rest of our discussion; the great bulk
of the validator's work is devoted to syntax checking.


>
> A difference was being pointed out between two quite different
> contexts. For me at least, there is but an analogy between html
> validation and the validation of an ordinary language argument.
> In the one case, a machine can do it, in the other a machine
> can't really do it well.

Yes.
The ability of HTML to deal with what is
going on 'semantically' is incompletely specified.
In the case of the truth functional logic, the semantics are
_provably_ complete.

My point was that the two beasts are of different species,
that 'semantics,' and 'interpretation,' as used in HTML
is different, and just sloppier, than the terms are when
used in logic, although there may be some similarities.
I notice IV also pointed this out.

It is a shame, but unavoidable, that terms like 'semantics'
are being subjected to ----- semantic drift! :)


> .............
> Why not take the opportunity to make a summary of these in a
> relatively easy to fathom form for the benefit of those who would
> appreciate a hand in this matter?
>

[Back to original message]