> HTML is like that. It's complex and huge. It might not
> be a bad idea (and a first level of achievement) if we
> could read back the same subset we write out, but the
> authors here are correct. HTML is HARD HARD HARD.
>
> (Just note that neither Microsoft nor Netscape are fully
> compliant after years of effort on the parts of multiple
> full-time programmers. And then there's the fact that the
> standard keeps changing in ways that make it incompatible
> with previous versions)
>
I don't think HTML is a good area for a wordprocessor. It and standard
wordprocessing have different metaphors [the page in HTML being infinite in
length, for one], and even different interpretations of the same code make
reading it a bear among browsers, let alone nonbrowser tools. Witness some
table tricks that choke Arachne but IE3 handles fine.... in such cases, who's
got the 'right' interpretation? However, how browsers handle HTML is a good
model for file formats, I think. In theory, they should ignore the stuff they
don't get... is that how AbiWord will work with its XML tags? Will I be able
to take the output from releases of AD 2040 and read them on 0.5.2?