Re: [Gzz] Encapsulating enfilade/string matching in CellTexter?

[b . fallenstein]

> > I don't agree. What setText() implies is that you can't give xanalogical
> > addresses for the text-- for example, when a user has pasted plain text
> into the
> > cell. In these cases, I believe the matching to be appropriate.
> 
> No, the opposite: you have no idea from where the user pasted the words
> so matching them and claiming that these are the words that mr. X wrote
> on 27.4.2004 is not right.

If the user pastes into a cell that had content before, and there is a
substantial match between what was there before and what is there after, I think
it is reasonable to assume that the user pasted an updated version of what
they originally pasted into that cell. If there is very little similarity, I
would in any case be for discarding that similarity (in extedit, if in a
1000-word text you find the word "the" both before and after extedit, it's not 
too
likely that the user removed all the rest and typed in different surroundings
instead).

> The point is that the content may come from a different cell, a different
> application, anything. Making a Xanalogical transclusion where you're
> not sure there is one is not right.
> 
> With ExtEdit you're pretty sure you know that the user *started* with
> the content and edited it so it's reasonably ok to say that it has
> something of the same...

I don't think the heuristic is much better. In my experience with Gzz <=
0.6, I've a number of times pasted different versions of the same text into the
same cell (e.g. text from a document ordinarily edited in a word processor),
but I can't recall ever having a need to paste a totally different text into
the same cell.

- Benja