>> Does that hold for programming language compilers as well? Does that
>> hold for UNIX shell scripts or, if fact, any formal textual way of
>> specifying to a computer what you want it to do?
> The problem is that there are a lot of HTML documents constructed by
> hand. It is not possible to correct all of these so the browsers
> have to be tolerant.
Not at the cost of breaking on correct HTML. Not *ever*!
And I still don't see what's wrong with providing good user
feedback. The errors are mostly easy to correct once they have been
pointed out.
Thus the fact that an HTML doc contains errors should always be
signalled to the user, and it should be easy for him to find out where
the browser thinks the error is.
See, I'm not against tolerance. It's the silent glossing over of
errors I disagree with.
> On the programming language issue, many programming language
> `errors' which are correctable are due to bad design in the first
> place. Many editors can insert missed ";" and remove spurious ","
> this implies that the grammars need not have resorted to this ugly
> kludge. In fact this is what we find. Occam, a formal language by
> anyone's standard gets by without a single {, } or ;.
Well,... er... except from the ;, so does HTML,... but that's beside
the point. :-)
Can you find me a programming language that doesn't need balanced
quotes, or some other way of terminating the quoted text? What happens
when the compiler/interpreter has to guess at where the programmer
meant to put the ending quote?
- Steinar