[xml-h] How semantic are links?

[Nik Barron]

> I think the biggest difference between a link created by an 
> author right in a document and another created by someone else as an 
> entry in a linkbase is that the former is so much easier to create. 

Easier to create, but potentially much harder to maintain over time. The web
is plagued by dead links, which is not a huge issue for general use but
could be catastrophic in a more critical environment (for example, using
online docs for safety critical process control etc). 

It's a case of cost of production vs cost of maintenance; in industry there
is a huge cost associated with keeping support information up to date (with
many product lifecycles measured in tens of years), and also hidden costs of
people using out of date information. Not all of this cost is down to
maintaining links of course, but a lot of information reuse is done manually
rather than with links because it's "easier" and "cheaper" to do at the
authoring stage this way.

Of course this isn't an either/or thing; you can mix 'vanilla'
unidirectional links with those from a linkbase by having a link proxy
service that processes the linkbase and inserts the appropriate HTML (or
whatever) markup. 

> Of course, that was then, and this is now.

And unfortunately the linking technology is still not as easily available as
it would be in an ideal world :-/ 

Just my random thoughts for the day!

Nik